Dai Z 2019

Developing Energy and Nutrient Mass Balances to
Inform Value Recovery Options in Municipal

Wastewater Treatment Systems
Thesis submitted to Newcastle University for the degree of Doctor of

Engineering in the Faculty of Science, Agricultural & Engineering
Author: Ziye Dai

Supervisors: Dr Adam Jarvis
Professor David Graham
Date: June 2019
Abstract
Municipal wastewater contains valuable substances such as energy, organic carbon and
nutrients. Recovery of these valuable substances may be a promising approach towards the
circular economy. However, their fate in the treatment processes is rarely studied holistically
and may have substantial variations between sites. The overall aim of this project is to reliably
determine the chemical energy and mass balances of wastewater treatment plants in the UK to
inform possible options for recovery of organic carbon (for energy), nitrogen and phosphorus.
This study firstly developed a novel method for rapidly measuring wastewater energy content
and determined the relationship between COD and chemical energy as 15.8 kJ/g COD. A
similar relationship is estimated for the sludge. Subsequently, chemical energy balance is built
on the COD balance. This study discovers only 20-30% of influent chemical energy can be
potentially recovered as CH4 for energy recovery. With this limited amount of energy recovered,
wastewater treatment is unable to be energy self-sustaining even that the influent energy is
more than 5-14 folds of the electricity consumption. Potential energy recovery opportunities
are found in the primary settlement, secondary treatment, sludge return liquor, and digested
sludge. In terms of nutrient, 17-41% of influent TN and 33-52% of influent TP are accessible
for recovery in the sludge return liquor and biosolid. The potential value recovery option is to
increase the capturing of nutrient into the sludge.
The impact of the implementation of energy and nutrient recovery measures on the overall
energy balance and mass balance are studied via modelling. The nutrient recovery measure
increases the nutrient loading in the sludge return liquor and biosolid and barely impacts the
energy balance. However, whilst reducing the electricity deficit, the energy recovery measures
tend to reduce the nutrient loading in the sludge return liquor but increase the proportion
distributed in biosolid and the final effluent.
The economic feasibility and sensitivity study of the implementation of the biosolid pyrolysis
is part of this research. Although it appears not to be economically viable at the current time,
it substantially improves the energy balance of the treatment work. Economically viability
could be achieved in the future if the influential capital cost can be reduced or better incentive
can be granted.
i
Declaration
I hereby declare that this thesis is my original research work. I truly acknowledged all the
sources of literature and information which has been used in this thesis. This thesis has not
been submitted for a degree at this or any other University previously.
Ziye Dai
15/03/2019
ii
Acknowledgements
I would like to acknowledge my primary supervisor Dr Adam Jarvis, with thanks for providing
me lots of academic and technical support and encouragement. I am very appreciated his
enormous patience and help throughout my whole EngD study, especially at the time I failed.
Many great thanks to Mrs. Justine Easten, who has always been kind and helpful and have been
looking after me in every aspect of this EngD study. I would also like to thank Dr Jaime
Amezaga for being such a supportive director of the STREAM centre in Newcastle University
I would like to thank STREAM IDC and my sponsoring company Northumbrian Water and
Scottish Water for offering me the funding to complete this EngD. I would like to thank Dr
Mark Haffey previously from Scottish Water and especially Mr. Andrew Moore from
Northumbrian Water, who have always been supportive and have provided me with all the
resources I needed for this project. I also would like to thank Mr. Luke Dennis, Mr. John
Robinson, Mr. Andrew Hinsmarsh, and Ms Laura Wilkinson from Northumbrian Water, and
Mr. Robert McGinty and Mr. Jamie Lynch for their assistance during the sampling work.
I would like to thank my colleagues and co-workers who have given me advice and help and
had fun time together. Particularly, I would like to thank Mr. Dave Race for all his help in the
laboratory. I would like to thank my STREAM cohort IV members for their care and love.
I am incredibly thankful to my parents, my families and my friends back in China. Many thanks
for their patience and love.
Last but not least, greatest gratitude to my wonderful wife Teng Zhang, who sacrifices a lot for
staying in the UK for being me accompany. Without her endless support, I may not be able to
get through all the dark moments in this journey.
iii
Table of Contents
Chapter 1 Introduction ............................................................................................................... 1
1.1 Background and drivers ................................................................................................... 1
1.1.1 Wastewater and wastewater treatment ...................................................................... 1
1.1.2 Drivers for value recovery from wastewater ............................................................ 2
1.1.3 Current status of value recovery from wastewater ................................................... 4
1.2 Aims and Objectives ........................................................................................................ 6
1.3 Thesis Structure ............................................................................................................... 8
Chapter 2 Literature Review ...................................................................................................... 9
2.1 The surrogate of chemical energy in wastewater and in sludge ...................................... 9
2.1.1 Surrogate of wastewater chemical energy .............................................................. 10
2.1.2 Surrogate of sludge chemical energy ...................................................................... 13
2.2 Fate of valuables in wastewater treatment work ............................................................ 13
2.2.1 Fate of COD, N and P in primary settlement .......................................................... 15
2.2.2 Fate of COD, N and P in secondary settlement ...................................................... 16
2.2.3 Fate of COD, N and P in sludge treatment ............................................................. 19
2.2.4 COD, N and P mass balance of wastewater treatment plants ................................. 20
2.3 Applying modelling of mass balance investigation and the impact of waste recovery . 21
2.3.1 Wastewater Mass balance Constraints .................................................................... 21
2.3.2 Applying modelling for mass balance construction ................................................ 22
2.3.3 Applying modelling for studying waste recovery ................................................... 23
2.4 Research Questions ........................................................................................................ 24
Chapter 3 Methodology ........................................................................................................... 26
3.1 Sample and data collection methods .............................................................................. 26
3.1.1 Wastewater and sludge sampling ............................................................................ 26
3.1.2 Sample storage ........................................................................................................ 27

3.1.3 Flow and energy consumption data collection ........................................................ 28
3.2 Chemical analysis .......................................................................................................... 29
3.2.1 Chemical analysis of wastewater ............................................................................ 29
3.2.2 Chemical analysis of sludge.................................................................................... 33
3.3 Statistical analysis .......................................................................................................... 37
3.4 Wastewater and sludge treatment modelling ................................................................. 38
3.4.1 Model construction ................................................................................................. 38
3.4.2 Influent characterization ......................................................................................... 46
3.4.3 Model calibration and validation ............................................................................ 53
3.5 Construction of mass and chemical energy balances ..................................................... 60
3.5.1 Calculation of mass loading .................................................................................... 60
3.5.2 Calculation of chemical energy loading.................................................................. 61
3.5.3 Flow balance ........................................................................................................... 62
3.5.4 Mass balance ........................................................................................................... 63
3.5.5 Chemical energy balance ........................................................................................ 65
3.5.6 Calculations related to the primary treatment ......................................................... 67
Chapter 4 Results and Discussion: Chemical energy of wastewater and sludge ..................... 70
4.1 Chemical energy of wastewater ..................................................................................... 70
4.1.1 Method Improvement.............................................................................................. 70
4.1.2 Wastewater characteristics ...................................................................................... 71
4.1.3 The relationship between COD and chemical energy content of wastewater ......... 73
4.1.4 The impact from diurnal variation to the chemical energy in wastewater .............. 80
4.2 Chemical energy of sludge ............................................................................................. 83
4.2.1 Sludge Characteristic .............................................................................................. 83
4.2.2 The relationship between solids and chemical energy content of sludge ............... 89
4.2.3 The relationship between COD and chemical energy content of sludge ................ 92
4.3 Chapter Summary .......................................................................................................... 94

Chapter 5 COD Mass Balance and Energy Balance of Wastewater Treatment Plants ............ 96
5.1 Model calibration and validation ................................................................................... 96
5.1.1 Wastewater and sludge treatment models ............................................................... 96
5.1.2 Calibration and validation of the wastewater treatment models ........................... 100
5.1.3 Calibration and validation of the sludge treatment models .................................. 103
5.2 The COD mass balance and chemical energy balance of treatment processes ............ 107
5.2.1 The COD mass balance and chemical energy balance of primary settlement units
........................................................................................................................................ 107
5.2.2 The COD mass balance and chemical energy balance of secondary biological
treatment (activated sludge) ........................................................................................... 110
5.2.3 The COD mass balance and chemical energy balance of sludge thickening and
dewatering facilities ....................................................................................................... 112
5.2.4 The COD mass balance and chemical energy balance of anaerobic digestion ..... 114
5.3 Energy recovery opportunities based on WWTP mass balances ................................. 116
5.3.1 Energy product of wastewater treatment .............................................................. 116
5.3.2 Energy product of sludge treatment ...................................................................... 117
5.3.3 Energy recovery potential from wastewater ......................................................... 119
5.4 Case studies of energy recovery .................................................................................. 126
5.4.1 Case study: Improving solids removal efficiency in primary settlement.............. 128
5.4.2 Case study: Reducing the sludge retention time prior to digestion ...................... 133
5.4.3 Case study: Recovering energy from the dewatered digested sludge via pyrolysis
........................................................................................................................................ 136
5.4.4 Case study: Replacing the activated sludge processes with an energy-generating
anaerobic membrane reactor unit ................................................................................... 139
5.5 Chapter summary ......................................................................................................... 143
Chapter 6 Results and Discussion: Nitrogen and Phosphorus Mass Balances of Wastewater
Treatment Plants .................................................................................................................... 145
6.1 Nitrogen and phosphorus mass balance of individual treatment processes ................. 145
6.1.1 Nitrogen and phosphorus mass balances of primary settlement units .................. 145
6.1.2 Nitrogen and phosphorus mass balance of secondary biological treatment as

activated sludge .............................................................................................................. 147
6.1.3 Nitrogen and phosphorus mass balance of sludge thickening and dewatering ..... 150
6.1.4 Nitrogen and phosphorus mass balance of anaerobic digestion ........................... 152
6.2 Discussion of the recovery opportunities based on plant wide mass balances ............ 152
6.2.1 Nutrient product of wastewater treatment............................................................. 152
6.2.2 Nutrient product of sludge treatment .................................................................... 155
6.2.3 Nutrient recovery potential from wastewater ....................................................... 157
6.3 Case studies of nutrient recovery ................................................................................. 160
6.3.1 Case study: Benefits to nutrient recovery of enhancing nutrient removal during
wastewater treatment ..................................................................................................... 161
6.3.2 Case study: Impact on nutrient recovery of improved primary settlement ........... 166
6.3.3 Case study: Impact on nutrient recovery of reducing sludge storage time ........... 169
6.3.4 Case study: Impact on nutrient recovery of using an anaerobic membrane reactor
as secondary biological treatment .................................................................................. 171
6.4 Chapter Summary ........................................................................................................ 173
Chapter 7 Business case study: pyrolysis of the biosolid for further energy recovery .......... 176
7.1 Introduction .................................................................................................................. 176
7.2 Method ......................................................................................................................... 177
7.2.1 The calculation of cash flows ............................................................................... 178
7.2.2 The calculation of payback period, NPV and IRR ................................................ 179
7.2.3 Sensitivity study on the uncertainty ...................................................................... 181
7.3 Result and Discussion .................................................................................................. 182
7.3.1 Energy balance ...................................................................................................... 182
7.3.2 Cash flow .............................................................................................................. 189
7.3.3 The economic evaluation and the sensitivity study on the uncertainties .............. 194
7.3.4 Environmental Benefit .......................................................................................... 200

7.4 Chapter Summary ........................................................................................................ 201
Chapter 8 Conclusion and Recommendation ......................................................................... 202
8.1 Conclusion ................................................................................................................... 202
8.1.1 Chemical energy of wastewater and sludge .......................................................... 202
8.1.2 Investigation of chemical energy balance and COD, nitrogen and phosphorus mass
balance ........................................................................................................................... 203
balance ........................................................................................................................... 206
8.2 Recommendation ......................................................................................................... 209
Appendix A. TN Adjustment for Sludge Sample .............................................................. 211
Appendix B. The value range of the kinetic and stoichiometry of the biological models. 212
Appendix C. Sludge density calculation ............................................................................ 217
Appendix D. Model calibration ......................................................................................... 219
Appendix E. The observed and predicted flow and concentration data ............................ 230
Appendix F. Conceptual modelling for investigating the impact of population equivalent to

energy and mass balance ........................................................................................................ 232
References .............................................................................................................................. 235

List of Figures
Figure 1-1 Schematic of typical wastewater treatment and sludge treatment ........................... 2
Figure 1-2 The schematic of linear economic model and circular economic model (IWA,
2016) .......................................................................................................................................... 3
Figure 3-1 Relationship between TP and COD of WWTP C................................................... 49
Figure 3-2 Relationship between TP and COD of WWTP C................................................... 49
Figure 4-1 Plots of Eww of different sample groups against COD, pCOD or TOC. (A) Eww of
Composite Samples against pCOD, (B) Eww of Composite Samples against TOC, (C) Eww of
Composite Raw against COD, (D) Eww of Composite Samples against COD; (E) Eww of Spot
Raw against COD, and (F) Eww of All Samples against COD ................................................. 76
Figure 4-2 Individual value plot of pCOD:COD of Composite Raw and Spot Raw ............... 81
Figure 4-3 The hourly flow, COD and energy loading to the WWTP A and C ....................... 82
Figure 4-4 Plots of Es against TSs, VSs and TOCs (A) Es against TSs, (B) Es against VSs, and
(C) Es against TOCs .................................................................................................................. 90
Figure 4-5 Physico-chemical fractionation of total COD in the sludge (Foladori et al., 2010)
.................................................................................................................................................. 92
Figure 4-6 The relationship between COD and VS in wastewater .......................................... 93
Figure 4-7 The relationship between energy content (kJ/g) and COD in sludge ..................... 94
Figure 5-1 The wastewater treatment model of WWTP A....................................................... 97
Figure 5-2 The wastewater treatment model of WWTP B....................................................... 97
Figure 5-3 The wastewater treatment model of WWTP C....................................................... 98
Figure 5-4 The wastewater treatment model of WWTP D ...................................................... 98
Figure 5-5 The model of WWTP C’s wastewater treatment and sludge thickening and
dewatering prior to anaerobic digestion ................................................................................. 100
Figure 5-6 The model of WWTP C’s anaerobic digestion and the subsequent sludge
dewatering .............................................................................................................................. 100
Figure 5-7 Comparison between the observed results and predicted results of the modelling of
four WWTPs (A)WWTP A,(B) WWTP B, (C)WWTP C and (D) WWTP D ........................ 102
Figure 5-8 Comparison between the observed results and predicted results of the modelling of
sludge treatment ..................................................................................................................... 106
Figure 5-9 The COD mass and chemical energy balances of primary settlement in a daily
basis at the four WWTPs investigated ................................................................................... 108
Figure 5-10 The impact of CODs concentration and fractions to the percentage COD captured
in primary sludge ................................................................................................................... 109
Figure 5-11 The impact of HRT to the percentage COD captured in primary sludge ........... 109
Figure 5-12 The COD mass and chemical energy balances of activated sludge in a daily basis
at the four WWTPs investigated ............................................................................................ 111
Figure 5-13 The TSS, VSS, COD mass balances and chemical energy balance of the sludge
thickening and dewatering processes in WWTP C ................................................................ 112
Figure 5-14 The COD and chemical energy balance of anaerobic digestion in WWTP C.... 114
Figure 5-15 The COD and chemical energy mass balance of the four WWTPs investigated116
Figure 5-16 The COD and chemical energy balance of the sludge treatment in WWTP C .. 118
Figure 5-17 Schematic illustration of different approaches to heat supply to the THP ......... 122
Figure 5-18 The breakdown of daily electricity consumption (kWh/d) of WWTP C ........... 127
Figure 5-19 The conceptual model CM-4-1 .......................................................................... 129
Figure 5-20 Schematic illustration of the dewatered digested sludge pyrolysis module ....... 137
Figure 5-21 The conceptual model CM-4-2 .......................................................................... 140
Figure 6-1 The nitrogen and phosphorus mass balances of primary settlement in a daily basis
at the four WWTPs investigated ............................................................................................ 146
Figure 6-2 The nitrogen and phosphorus mass balances of secondary biological treatment in a
daily basis at the four WWTPs investigated .......................................................................... 148
Figure 6-3 The nitrogen and phosphorus mass balances of the sludge thickening and
dewatering processes in WWTP C ......................................................................................... 151
Figure 6-4 The nitrogen and phosphorus balances of the four WWTPs investigated ........... 153
Figure 6-5 The nutrient balance for the sludge treatment process of WWTP C .................... 156
Figure 6-6 The CM-5-1 model ............................................................................................... 162
Figure 7-1 The energy balance (EBGtG) of the Base Case ..................................................... 184
Figure 7-2 The energy balance (EBGtG+FP) of the Base Case system with biosolid pyrolysis
................................................................................................................................................ 186
Figure 7-3 The energy balance (EBCHP) of the system has no CH4 grid injection ................ 188
Figure 7-4 The energy balance (EBCHP+FP) of the system has no CH4 grid injection but with
biosolid pyrolysis ................................................................................................................... 189
Figure 7-5 The percentage breakdown of income of different cash flow .............................. 194
Figure 7-6 The spider chart of payback period ...................................................................... 194
Figure 7-7 The spider chart of the NPV ................................................................................. 195
Figure 7-8 The spider chart of the IRR .................................................................................. 195
Figure 7-9 The probabilistic distribution of NPV obtained from Monte Carlo Analysis
(Korytárová and Pospíšilová, 2015) ...................................................................................... 198
Figure F-1 The conceptual model CM-F-1 ............................................................................ 232
List of Tables
Table 1-1 Overview of the 4 Investigated WWTPs ................................................................... 7

Table 2-1 The kJ/g COD values of commonly found organics in wastewater studied by Owen
(1982) ....................................................................................................................................... 11
Table 2.2 The ratio of COD, TN and TP loading in the sludge return liquor to the influent
loading reported in literatures .................................................................................................. 19
Table 3-1 The default and adjusted values of related kinetic of the Al2(SO4)3·18H2O for
virtual sCOD removal .............................................................................................................. 40
Table 3-2 The configuration of the primary settlement tanks of each WWTP ........................ 42
Table 3-3 The configuration of the activated sludge process and the subsequent clarification
process for each WWTP .......................................................................................................... 43
Table 3-4 The volume of the sludge storage tanks................................................................... 45
Table 3-5 The concentration and fraction used for characterizing the imported sludge .......... 52
Table 3-6 The influent liquid streams, effluent wastewater streams and sludge streams of the
flow balance of each individual process .................................................................................. 63
Table 3-7 The influent and effluent wastewater streams and sludge streams used in the mass
balance of each individual process .......................................................................................... 64
Table 3-8 The influent and effluent wastewater streams and sludge streams of the energy
balance for each individual process ......................................................................................... 66
Table 4-1 Mean COD and Eww (± standard deviation) of wastewater samples from this
research compared to literature values ..................................................................................... 72
Table 4-2, Best regression for estimating wastewater energy .................................................. 74
Table 4-3 Regression equations of regressions between Eww and pCOD, TOC and COD
shown in Figure 4-1 ................................................................................................................. 76
Table 4-4. Common organic compounds and their calculated energy values .......................... 77
Table 4-5 Chemical energy contributed by commonly found substances unaccounted for in
the COD measurement ............................................................................................................. 79
Table 4-6 The mean value (± standard deviation) of TS%, VS/TS%, TOC/TS%, TN/TS%,
TP/TS% and chemical energy content of the dried sludge and wet sludge ............................... 1
Table 4-7 Best regression for estimating wastewater energy ................................................... 89
Table 4-8 Regression equations of regressions between Es and TSs, VSs and TOCs shown in
Figure 4-4 ................................................................................................................................. 90
Table 4-9 Energy related characteristic of sludge reported in literature .................................. 91
Table 5-1 The influent water quality, and other input variables, used for the modelling of each
WWTP ..................................................................................................................................... 99
Table 5-2 The predicted different type of CODs of the feed sludge to different sludge
thickening and dewatering ..................................................................................................... 105
Table 5-3 The predicted influent and effluent CODs concertation and the CODs removal
efficiency of primary settlement in four WWTPs .................................................................. 109
Table 5-4 The predicted sCOD and COD concentration in different filtrates and centrates . 113
Table 5-5 The estimated chemical energy production potential from the four study WWTPs
................................................................................................................................................ 120
Table 5-6 The estimated electricity production of different wastewater................................ 124
Table 5-7 The per unit electricity consumption of the primary settlement, aeration, sludge
treatment and fixed consumption of the WWTP C ................................................................ 128
Table 5-8 The chemical energy balance of the models with different SREPS ........................ 131
Table 5-9 COD, Substrate COD and Heterotrophic biomass COD of the sludge fed to the raw
sludge storage tank ................................................................................................................. 132
Table 5-10 The electricity balance of the models with different SREPS ................................ 133
Table 5-11 The change of sCOD, pCOD and COD over different raw sludge storage time . 135
Table 5-12 The chemical energy balance of models with different raw sludge storage time 135
Table 5-13 The electricity balance of models with different Raw sludge storage time ......... 136
Table 5-14 The chemical energy balance of the baseline model outputs and the AnMBR
model outputs ......................................................................................................................... 141
Table 5-15 The electricity balance of the activated sludge and the AnMBR processes ........ 142
Table 6-1 The particulate phosphorus, TP, the particulate nitrogen and TN of the raw
wastewater of the four study WWTPs ................................................................................... 147
Table 6-2 The predicted concentrations of different types of nitrogen of the activated sludge
treated effluent at the four WWTPs investigated ................................................................... 150
Table 6-3 Nutrient concentration of different filtrates or centrates ....................................... 152
Table 6-4 The estimated percentage TN and TP kept in the dewatered digested sludge ....... 157
Table 6-5 The estimated percentage TN and TP (existing as NH4+-N and PO43-_P) kept in the
raw sludge centrate and digested sludge centrate .................................................................. 159
Table 6-6 The TN and TP percentage mass balance of the baseline, ferric dosing and EBPR
models .................................................................................................................................... 163
Table 6-7 The concentration and loading of NH4+-N and PO43--P in the raw sludge centrate
and digested sludge centrate of models with different phosphate removal scenarios ........... 164
Table 6-8 The TN distribution in the output of the combined wastewater and sludge treatment
in models with different phosphate removal scenarios .......................................................... 164
Table 6-9 The TP distribution in the output of the combined wastewater and sludge treatment
in models with different phosphate removal scenarios .......................................................... 164
Table 6-10 The chemical energy balance of the models with different phosphate removal
scenarios ................................................................................................................................. 165
Table 6-11 The electricity balance of the models with different phosphate removal scenarios
................................................................................................................................................ 166
Table 6-12 The concentration and loading NH4+-N and PO43--P of raw sludge centrate and
digested sludge centrate under different solid removal efficiency of primary settlement ..... 167
Table 6-13 The TN distribution in the output of the combined wastewater and sludge
treatment under different solid removal efficiency of primary settlement ............................ 168
Table 6-14 The TP distribution in the output of the combined wastewater and sludge
treatment under different solid removal efficiency of primary settlement ............................ 168
digested sludge centrate under different raw sludge tank volume scenarios ......................... 170
treatment under different raw sludge tank volume scenarios ................................................ 170
treatment under different raw sludge tank volume scenarios ................................................ 171
digested sludge centrate of models with different secondary biological treatment ............... 172
treatment in different scenarios of secondary biological treatment ....................................... 172
treatment of models with different secondary biological treatment ...................................... 173
Table 7-1 The estimated CAPEX of the components of the biosolid pyrolysis .................... 190
Table 7-2 The annual cash flow of different energy balance and under different scenarios .. 193
Table 7-3 The impact to the NPV, IRR and Payback period if a ±20% change is given to the
parameters tested for sensitivity............................................................................................. 196
Table A-1 Adjusted TN% of different type of sludge of Howdon WWTP ............................ 211
Table B-1 The value range of the kinetic and stoichiometry of the biological models ......... 212
Table D-1 The calibrated parameters of the wastewater treatment model of WWTP A ........ 219
Table D-2 The calibrated parameters of the wastewater treatment model of WWTP B ........ 220
Table D-3 The calibrated kinetics and stoichiometry of the wastewater treatment model of
WWTP C ................................................................................................................................ 221
Table D-4 The calibrated kinetics and stoichiometry of the wastewater treatment model of
WWTP D................................................................................................................................ 222
Table D-5 The calibrated kinetics and stoichiometry of the primary sludge storage, thickening
and dewatering ....................................................................................................................... 223
Table D-6 The calibrated parameters of the SAS storage, thickening and dewatering ......... 223
Table D-7 The calibrated parameters of the raw sludge storage, thickening and dewatering 224
Table D-8 The calibrated parameters of the anaerobic digestion process and subsequent
sludge storage, thickening and dewatering ............................................................................ 224
Table E-1 The observed and predicted flow and concentration of the modelling of wastewater
treatment of four WWTPs ...................................................................................................... 230
Table E-2 The observed and predicted flow and concentration of the modelling of sludge
thickening, dewatering and digestion of WWTP C ............................................................... 231
Table F-1 Dimensions of the primary settler, aeration tank and final clarifier ...................... 233
Table F-2 Chemical energy balance and COD, TN and TP mass balance of the conceptual
treatment work in different flow condition ............................................................................ 234
Chapter 1 Introduction
This Chapter first introduces the background and drivers for the research undertaken in Section
1.1. The aims and objective are proposed in Section 1.2. The thesis structure is shown in Section
1.3.
1.1 Background and drivers
1.1.1 Wastewater and wastewater treatment
Municipal wastewater comes from the water from toilets, baths, sinks and washing machines
in homes, from water used in industry, and from rainwater runoff (DEFRA, 2012). This water
is contaminated with organic matter, nutrients, metals, oils, pathogens, and other pollutants
(DEFRA, 2012, Tchobanoglous et al., 2003). Direct discharges of wastewater to the
environment may cause severe damage to the water environment, such as eutrophication of the
receiving waters caused from the excessive input of nutrients, ecosystem damage to receiving
waters due to the oxygen depletion during microbial biodegradation of organic matter, and
health risks caused by the pathogens from wastewater entering into waters used for recreational
purposes e.g. bathing waters (DEFRA, 2012). Therefore, wastewater treatment is crucial to
ensure that wastewater quality is of a suitable standard to return to the environment.
Wastewater from different sources is collected in the sewerage network and flows into
wastewater treatment works. Wastewater treatment generally has three stages: (1) preliminary
treatment (2) primary treatment and (3) secondary treatment (Tchobanoglous et al., 2003,
DEFRA, 2012). Preliminary treatment typically involves screening out large solids, such as
wood and plastics, and removal of grit. Primary treatment has the objective of settling
suspended particulate matter. Secondary treatment is usually biological and is used to remove
the residual particulate and soluble contaminants. In some cases tertiary treatment is required
for removing specific types of pollutant (Tchobanoglous et al., 2003). Once pollutant
concentrations are lowered to below the consent level, wastewater can be discharged back to
environment.
1
The treatment processes, especially the primary and secondary treatment, produce sludge as a
by-product. In current practise the produced sludge is commonly sent for anaerobic digestion
(AD). Since AD requires the feed sludge to have a certain moisture level (Mills et al., 2014),
sludge goes through a thickening process to remove the extra undesired water content. The
anaerobic digestion of sludge produced digested sludge and biogas contains biomethane and
CO2 (Smyth et al., 2016, Wan et al., 2016, Tchobanoglous et al., 2003). The digested sludge
will be dewatered and usually sent for land application (DEFRA, 2012). The produced
biomethane will be commonly used onsite for electricity generation or injected to the gas grid
(Mills et al., 2014). A schematic flow diagram of the wastewater and sludge treatment process
is shown in Figure 1-1.
Raw Preliminary Primary Secondary Final

Wastewater Settlement Settlement Treatment Effluent
Primary Secondary
Sludge Sludge
Sludge
Reject Water Thickening
from Sludge Thickening
Thickened Sludge
Anaerobic Bio-methane
Digestion and CO2
Digested
Sludge
Sludge
Reject Water Dewatering
from Sludge Dewatering
Dewatered
Digested
Figure 1-1 Schematic of typical wastewater treatment and sludge treatment
1.1.2 Drivers for value recovery from wastewater
In essence, wastewater is a used resource that is disposed of to a wastewater treatment facility

in order to minimize the harm of that disposal. This reflects the conventional economic model
of “take-make-consume-dispose”: resources are taken from the environment, made into
products, consumed by society, and finally disposed back to the environment (IWA, 2016).
This economic model heavily relies on the abundance of natural resources (IWA, 2016). If the
2
natural resource is finite, then this economic model is not sustainable (IWA, 2016). A schematic
illustration of the linear economy model is shown in Figure 1-2.
Linear Economic Model
Take Make Consume Dispose
Natural Waste
Resource
Circular Economic Model
Take
Natural
Resource
Recover/
Recycle/ Make
Reuse
Consume
Waste
Figure 1-2 The schematic of linear economic model and circular economic model (IWA, 2016)
Pearce and Turner (1990) proposed a concept of circular economy. In 2015, the European
Commission defined the circular economy as follows: “In a circular economy, the value of
products and materials is maintained for as long as possible. Waste and resource use are
minimised, and when a product reaches the end of its life, it is used again to create further
value.” The European Union also set a target to recycle 65% of municipal waste by 2035
(European Commission, 2019). Wastewater is rich in organic carbon, nitrogen and phosphorus
content (Ainger et al., 2009). Cellulose, plastics, fertilizers, energy of methane, hydrogen and
electricity have all been reported to be recoverable from wastewater (van der Hoek et al., 2016,
Scherson and Criddle, 2014, McCarty et al., 2011, Cotterill et al., 2018, van der Hoek et al.,
2018). Therefore, wastewater is no longer just a waste to dispose of, but actually a resource to
3
society. Wastewater treatment works can then be transformed into a chemical process plant for
extracting these valuable substances back into products (Ainger et al., 2009). The recovery of
chemical energy, nitrogen and phosphorus from wastewater could also bring in environmental
benefit, such as tackling the fossil scarcity and the phosphorus shortage (Mayer et al., 2016,
Wan et al., 2016). Value recovery from wastewater will therefore play an important role in the
circular economy (IWA, 2016).
1.1.3 Current status of value recovery from wastewater
Substances commonly recovered from wastewater are organic carbon, for chemical energy, and
nutrients for fertilizer (McCarty et al., 2011, Mayer et al., 2016, Stoll et al., 2018). Currently
full-scale value recovery is mainly from wastewater treatment sludge rather than directly from
wastewater (Bowen et al., 2010, Smyth et al., 2016, Mills et al., 2014, Münch and Barr, 2001).
The approach to energy recovery is anaerobic digestion of the sludge for methane production
(Smyth et al., 2016, Bowen et al., 2010, Mills et al., 2011). The approach to nutrient recovery
is typically to utilize the digested sludge as fertilizer for agricultural land application.
Alternatively phosphate and ammonium may be precipitated to produce the slow release
fertilizer struvite (magnesium ammonium phosphate), usually from the reject water produced
from digested sludge dewatering (Münch and Barr, 2001, Ostara, 2010, Paz-Ferreiro et al.,
2018).
However, current practices only recover limited amounts of the valuable substances flowing
into wastewater treatment plant (WWTP). A mass balance study of a Singapore WWTP
suggested that only approximately 18% of the organic carbon from the raw wastewater was
recovered as methane (Shi, 2011). Meanwhile, 11% of the nitrogen and 20% of the phosphorus
from the raw wastewater were found in the reject water produced from the digested sludge
dewatering, whilst 12% of nitrogen and 43% of phosphorus from the raw wastewater ended in
the dewatered digested sludge (Shi, 2011). It is worth noting that the recovery efficiency can
vary substantially from site to site. In another mass balance study of an Austrian WWTP,
approximately 36% the organic carbon from the raw wastewater was recovered as methane
(Wett et al., 2007).
4
Mass and energy balance is a very useful technique for process evaluation, designing and
benchmarking WWTP (Yeshi et al., 2013). More importantly, it can be used to guide process
improvement to promote value recovery from wastewater (Yeshi et al., 2013). However, there
are few studies that perform plant wide holistic mass balance investigation of organic carbon,
nitrogen and phosphorus together (Yeshi et al., 2013). Meanwhile, the construction of mass and
energy balance required proper flow and concentration measurement of each influent and
effluent of the treatment process (Tchobanoglous et al., 2003). The construction could be
constraint by the poor flow and concentration monitoring, since water companies do not often
synchronously measure flow and concentration on each influent and effluent of their treatment
process (personal communication, M Haffey, Scottish Water, April 2015, personal
communication, L Wilkinson, Northumbrian Water, June 2015).
As mentioned above the organic carbon recovered from wastewater is used for energy
application. Although some efforts have been made to construct chemical energy balances for
wastewater treatment works from the organic carbon (expressed as chemical oxygen demand,
COD) mass balance, the relationship between chemical energy and the COD of wastewater is
yet to be reliably defined (Wan et al., 2016, Wett et al., 2007). The approach to measurement
of chemical energy is to dry the wastewater first, and then to measure the calorific value of the
dried wastewater via bomb calorimetry (Korth et al., 2017, Shizas and Bagley, 2004, Heidrich
et al., 2010). The drying stage is vital to the accuracy of the energy measurement because it has
to retain as much of the energetic organic carbon in the dried wastewater and it should be time
efficient to allow large amount of measurements (Heidrich et al., 2010, Korth et al., 2017). The
current method applied for drying is freeze-drying. However, whilst this is able to retain more
than 80% of the organic carbon originally contained in the wastewater it requires approximately
4 weeks to process a sample (Heidrich et al., 2010). Furthermore, neither freeze driers or bomb
calorimeters are equipment commonly possessed by water companies (Korth et al., 2017,
Schaum et al., 2016). Therefore, to construct a reliable chemical energy balance, a reliable and
time efficient drying method is required. With greater amount of sample being analyzed, the
relationship between chemical energy and COD or another proper surrogates can be better
defined.
5
The implementation of value recovery initiatives may have a negative impact on existing
treatment performance at WWTPs. As stated in Section 1.1.1 and shown Figure 1-1, wastewater
and sludge treatment consists of a series of mechanical and biological processes. The
introduction of new value recovery technologies, or the improvement of current value recovery
technologies at a WWTP, may influence the performance of a process and consequently affect
the downstream processes. For instance, if a WWTP has a nutrient discharge consent, the
recirculation of its nutrient rich reject water from the digested sludge dewatering back to head
of the work, will stress the treatment process (Münch and Barr, 2001). Implementation of
struvite recovery, which recovers the nutrient from rejected water (from the sludge dewatering)
into fertilizers, will reduce the stress to the wastewater treatment performance (Münch and Barr,
2001). However, in the example of cellulose recovery, if the cellulose, which is an organic
carbon resource, is captured prior or post the primary treatment but not sent for digestion, the
organic loading carbon loading to the digester will be reduced (van der Hoek et al., 2016). This
causes an adverse effect to the current organic carbon recovery as biomethane production
becomes smaller (van der Hoek et al., 2016).
1.2 Aims and Objectives
Current efforts to recover value from WWTPs focus on sludge, but the degree to which
substances are recovered is variable and/or low (Shi, 2011, Wett et al., 2007). Furthermore,
recovery of valuable substances directly from wastewater is currently limited. In this research
the following aspects of value recovery were therefore explored:
• The potential for recovery of value directly from municipal wastewaters

• Improving the efficiency of current value recovery approaches used on sludge
The foundation for delivering such improvements in value recovery is a clear quantitative
understanding of the chemical and energy mass balances at WWTPs, such that there is a proper
appreciation of the fate of potentially valuable organic carbon, nitrogen and phosphorus during
wastewater and sludge treatment. Furthermore, knowledge of how these mass balances vary
between WWTPs is required. Only then can the benefits of value recovery options be truly
evaluated.
6
Therefore, the overall aim of this project is to reliably determine the chemical energy and mass
balances of wastewater treatment plants in the UK to inform possible options for recovery of
organic carbon (for energy), nitrogen and phosphorus.
To fulfil this overall aim, the objectives were set, as follows:

1. Develop a reliable and practical method for determination of the chemical energy
content of wastewater, and identify a reliable surrogate of the chemical energy in
wastewater and in sludge
2. Construct chemical energy and mass balances for 4 WWTPs to determine the fate of
chemical energy, organic carbon (expressed as COD), nitrogen and phosphorus in the
wastewater and sludge treatment (via modelling)
3. Identify potential waste recovery opportunities and explore the impact of their
implementation on the existing treatment process (via modelling)
In order to obtain representative findings, 4 WWTPs from the sponsor companies of Scottish
Water and Northumbrian Water were chosen in this study. These 4 WWTPs were intentionally
selected as they are varied in terms of population equivalent (P.E.), wastewater treatment
processes, and sludge treatment processes. Details for each of them are shown in Table 1-1.
Table 1-1 Overview of the 4 Investigated WWTPs

Population
WWTPs Wastewater Treatment Process Sludge Treatment Process
Equivalent
Primary Settlement with Chemical Dosing for
WWTP A 40,000 Phosphate removal and Activated Sludge with N/A
Nitrification
Sludge Thickening and
WWTP B 230,000 Primary Settlement and Activated Sludge
Dewatering
Sludge Thickening,
WWTP C 1,000,000 Primary Settlement and Activated Sludge Dewatering and Advanced
Anaerobic Digestion
Primary Settlement and Activated Sludge with
WWTP D 180,000 N/A
Nitrification
7
1.3 Thesis Structure
The remaining 7 chapters of this thesis are arranged as follows:
Chapter 2 - A literature review focusing on the determination of surrogate of chemical energy

in wastewater and in sludge and the study of the fate of the valuable organic carbon, nitrogen
and phosphorus in the wastewater and sludge treatment processes
Chapter 3 - Field and analytical methods used, the procedure of wastewater and sludge
treatment modelling, and the construction of chemical and energy balances
Chapter 4 - Results and Discussion of the method developed for determination of chemical
energy in wastewater, and discussion of the surrogates for estimating the chemical energy in
wastewater and sludge
Chapter 5 – Results and Discussion of chemical energy and COD mass balance, and discussion
of potential chemical energy recovery opportunities from wastewater and the impact of their
implementation
Chapter 6 - Results and Discussion of nitrogen and phosphorus mass balance, and discussion
of potential nutrient recovery opportunities from wastewater and the impact of their
implementation
Chapter 7 – Business case study on biosolid gasification for energy recovery, which includes
discussion of the economic evaluation and sensitivity studies on the uncertainties
Chapter 8 – Conclusions and Recommendation
8
Chapter 2 Literature Review
This chapter first reviews the surrogate of chemical energy in wastewater and in sludge in
Section 2.1. It then assesses the fate of valuable organic carbon, nitrogen and phosphorus in
the wastewater and sludge treatment processes in Section 2.2. It reviews application of
modelling in the mass balance investigation and in studying the impact of process
reconfiguration in Section 2.3. The research questions are then proposed in Section 2.4.
2.1 The surrogate of chemical energy in wastewater and in sludge
Chemical energy contained in wastewater is estimated approximately 5 times greater than

electricity consumed for during treatment (Wan et al., 2016). Wastewater also contains thermal
and kinetic energy (McCarty et al., 2011). Wastewater treatment has the potential to become an
energy net producer (McCarty et al., 2011). But, in reality it is the opposite as the chemical
energy within wastewater is not used in its full potential. It is worth, then to know the fate of
different types of energy during the wastewater.
Recovery of thermal energy requires heat pump which consumes electricity. Alternative fuels
for heating, such as natural gas or fuel oil, are equally expensive as electricity, then thermal
recovery is costly(McCarty et al., 2011). Kinetic energy is caused by the movement of
wastewater and hence it can be recovered by hydropower electricity generation (Power et al.,
2014). Kinetic energy recovery has a strict requirement: flow rate of wastewater. Power et al.
(2014) reported that only 14 out of 100 Irish wastewater treatment plants are suitable for
hydropower electricity generation, and those plants are relatively large. Also, thermal and
kinetic energy recovery are rarely conducted.
Currently, energy recovery focus on the chemical energy capture from wastewater and sludge
(McCarty et al., 2011, Stoll et al., 2018, Scherson and Criddle, 2014). Therefore, it is useful to
understand the chemical energy balance of the wastewater and sludge treatment to discover the
potential chemical energy reserve.
9
The experimental measurement of the chemical energy content of sludge and wastewater is
commonly done via bomb calorimetry (Shizas and Bagley, 2004, Heidrich et al., 2010, Korth
et al., 2017, Smyth et al., 2016, Schaum et al., 2016). Because more than 70% of the mass in
both wastewater and sludge is from non-combustible water, the wastewater and sludge will be
dried first and then analyzed for chemical energy content via bomb calorimetry (Korth et al.,
2017, Smyth et al., 2016, Shizas and Bagley, 2004, Heidrich et al., 2010, Schaum et al., 2016).
However, bomb calorimetry is not commonly possessed in WWTPs due to its cost and
complicated set up (for example, it requires oxygen cylinders) (Schaum et al., 2016, Smyth et
al., 2016). The chemical energy content of the wastewater and sludge is therefore not able to
be routinely measured (Smyth et al., 2016, Schaum et al., 2016). Chemical energy balance is
then calculated on the mass balance of a chosen surrogate of wastewater and sludge (Mills et
al., 2014, Wan et al., 2016, Wett et al., 2007). The determination of the relationship between
chemical energy content and the suitable surrogate of it is crucial to the accuracy of the
chemical energy balance.
2.1.1 Surrogate of wastewater chemical energy
Since chemical energy is mainly stored in the organic contaminants and chemical oxygen
demand (COD) is commonly used to represent the concentration of the organic contaminant in
wastewater, COD is the preferable surrogate of the chemical energy content in wastewater
(Owen, 1982, Korth et al., 2017, Scherson and Criddle, 2014, Tchobanoglous et al., 2003).
Owen (1982) firstly proposed the concept of using COD to estimate the energy content of
wastewater. The study measured the calorific value of several organic compounds that may
exist in wastewater. The oxygen demand of the complete combustion (CODCombustion) per gram
of those substances were also studied based on both theoretical enthalpy calculation and
experimental measurement. It was obtained by diving calorific value by the CODCombustion,
chemical energy within per gram of COD. The kJ/g COD of the measured organic ranged from
12.6 to 15.5 and the mean value was 14.2 (shown in Table 2-1). The study chose to use the
value of 13.9 kJ/g COD for its energy-related calculation. In this case, the energy calculation
used by Owen (1982) is based on BTU/lb. 13.9 kJ/g is equivalent to a round up number of
6,000 BTU/lb while 14.2 kJ/g is equivalent to 6,100 BTU/lb. This 13.9 kJ/g COD value was a
10
reasonable assumption, but it was not selected based on the proportion of the concentration of
organic compound in the wastewater.
Table 2-1 The kJ/g COD values of commonly found organics in wastewater studied by Owen (1982)
Name Chemical Formula Measured kJ/g COD Calculated kJ/g COD

Wastewater Organic C10H18O3N 15.5 13.9
Bacterial Cells C5H7O2N 14.7 15.4
Carbohydrates C6H10O5 14.7 12.7
Fats C18H3 6O2 15.1 14.5
Proteins C71H128O23S N/A 14.1
Acetic Acid C2H4O2 13.6 12.6
Benzene C6H6 13.6 13.8
Benzoic Acid C7H6O2 13.5 13.0
Butyric Acid C4H8O2 13.7 13.8
Carbon Dioxide CO2 0 0
Ethanol C2H6O 14.3 14.5
Furfural C5H4O2 14.7 12.8
Hydrogen Sulfide H2S N/A N/A
Lactic Acid C3H6O3 14.1 12.6
Levulinic Acid C5H8O3 N/A 13.2
Methane CH4 13.9 15.4
Methanol CH4O 15.2 14.4
Phenol C6H6O 13.7 13.5
Propionic Acid C3H6O2 13.7 13.6
Vanilin C8H8O3 14.0 13.0
Shizas and Bagley (2004) conducted the first study to experimentally measure chemical energy
content in wastewater. In this research, wastewater was oven dried at 104oC to remove all the
moisture. Subsequently, the dried wastewater was fed to the bomb calorimeter for chemical
energy measurement. The study determined a result of 14.7 kJ/g COD, which is higher than the
value that Owen (1982) reported. The 14.7 kJ/g COD was estimated based on one raw
wastewater sample which could not be considered representative.
Heidrich et al. (2010) also measured chemical energy content in wastewater. The study
followed a similar approach by drying the wastewater sample first, and subsequently measuring
11
its energy content via bomb calorimetry. However, Heidrich et al. (2010) used freeze drying
instead of the oven drying method. The study showed that freeze drying was able to retain 73-
81% of COD in the dried residue compared to oven drying which only retained 51-56%. Thus,
this leads to a higher measured value of chemical energy content of wastewater and a higher
value kJ/g COD. In this case, the study experimentally measured two raw wastewater samples
and determined values of 17.7 to 28.7 kJ/g COD respectively.
Rather than reporting energy content as kJ/g COD, Korth et al. (2017) used linear regression
to determine the mathematical relationship between chemical energy content of wastewater
and COD. His experiments concluded that chemical energy content in wastewater (kJ/L) equal
to 3.883×COD (g/L) +2.045. The equation was determined by applying 20 data points from
which 1 were from data point from Shizas and Bagley (2004) and 2 were from Heidrich et al.
(2010). However, whilst acknowledging that freeze-drying is a preferable method for
wastewater energy measurement, Korth et al. (2017) nevertheless used oven drying for the
majority of samples, because of the long processing time for freeze drying. The r2 of the linear
regression is reported as 0.65 (Korth et al., 2017).
The relationship between chemical energy and COD has only been experimentally studied on
20 raw wastewater samples globally. The experimental measured values reported Shizas and
Bagley (2004) and Heidrich et al. (2010) vary by up to 95 % (comparing to the 14.7 KJ/g COD).
Meanwhile, Korth et al. (2017) did suggest using linear regression to estimate the chemical
energy content of wastewater from COD, but the majority of the data used for linear regression
analysis were obtained from the oven drying method which is proven not effectively in
retaining COD. Therefore, the relationship between COD and chemical energy was poorly
defined and needs to be reliably determined.
Unlike the COD, N and P measurement, which can be conducted on aqueous samples and
completed within hours or days, the chemical energy measurement requires dried sample and
weeks to complete (Korth et al., 2017, Heidrich et al., 2010). The drying stage is time
consuming, as freeze drying takes 4 weeks to dry enough wastewater to generate approximately
0.5 g of dried residue to be used in the bomb calorimetry (Heidrich et al., 2010). The time-
consuming drying process hinders the large amount experimental measurement on chemical
12
energy content of wastewater and the subsequent study of the relationship between chemical
energy and COD.
2.1.2 Surrogate of sludge chemical energy
Total solid (TS) content is usually used as a surrogate of sludge chemical energy (Mills et al.,
2014, Shizas and Bagley, 2004, Schaum et al., 2016, Smyth et al., 2016). The chemical energy
content of the sludge is calculated by multiplying the calorific value of the dried solid content
with TS content (Mills et al., 2014). This is because the TS content is routinely measured in
sludge treatment and is commonly used to describe the performance and mass balance of the
sludge treatment (Smyth et al., 2016, Bowen et al., 2010).
Unlike wastewater, hundreds of sludge samples have been studied for chemical energy content
of the TS content (Schaum et al., 2016). The result does vary from types of sludge and studies.
The reported chemical energy content of TS content ranges from 15.9 to 17.9 kJ/g for primary
sludge, 12.2 to 18.1 kJ/g for surplus activated sludge, and 11.1 to 15.7 kJ/g for digested sludge,
respectively (Schaum et al., 2016, Shizas and Bagley, 2004, Smyth et al., 2016). There is a
considerable variation in the chemical energy content of TS content in the sludge, especially
for the surplus activated sludge and digested sludge.
Moreover, Shizas and Bagley (2004), Schaum et al. (2016) and Smyth et al. (2016) all
suggested that the sludge chemical energy is also related to volatile solid (VS) content because
most of the energetic organic is contained in the VS. The chemical energy content of VS content
reported in these three studies range from 21.0 to 26.6 kJ/g for primary sludge, 20.6 to 21.3
kJ/g for surplus activated sludge, and 20.5 to 30.7 kJ/g for digested sludge, respectively
(Schaum et al., 2016, Shizas and Bagley, 2004, Smyth et al., 2016). There is also a considerable
variation in the chemical energy content of VS content in the sludge. Therefore, a validation of
the relationship between the chemical energy and the TS and VS content of sludge is required.
2.2 Fate of valuables in wastewater treatment work
Naden et al. (2016) estimated 1,460,000 tonnes organic carbon, 263,000 tonnes of nitrogen and
43,600 tonnes of phosphorus flowed through United Kingdom’s WWTPs in 2010. Using a ratio
13
of chemical oxygen demand (COD) to total organic carbon as 3:1 (Henze and Comeau, 2008),
4,380,000 tonnes COD flows into the treatment work. Reportedly, 1 gram of COD contains
13.9 kJ of chemical energy (Owen, 1982). Then, an estimated 16,900 GWh of chemical energy
flows into the UK’s wastewater treatment. Wastewater treatment uses up to 1% of the national
electricity (POST, 2007). According to the Historical electricity data: 1920 to 2017 published
by Department for business, Energy & Industrial Strategy, the annual electricity production of
the UK in 2010 was 333,205 GWh. This suggests that electricity consumed for wastewater
treatment is approximately 3,332 GWh electricity per annum. The energy reserve in the
wastewater is approximately 5 times greater than the energy used for treatment purposes. For
nutrient, over the same period (2010), the UK’s annual fertilizer consumption was
approximately 1,000,000 tonnes of nitrogen and 84,000 of phosphorus (AIC, 2011). Indicating,
that more than a quarter of the nutrient demand for agricultural consumption can be found in
wastewater.
At present, these valuables are not fully recovered into usable resource from wastewater
treatment plants. Actually, part of them are converted and emitted into the atmosphere, another
portion will be captured in the sludge and ultimately released to land, and only a small part of
it is recovered into the products (i.e. CH4 or fertiliser of struvite) that we can use (Mills et al.,
2014, Münch and Barr, 2001, Tchobanoglous et al., 2003, Wett et al., 2007). Besides, part of
these valuable products will discharged back to surface water bodies. Naden et al. (2016)
estimated that in 2010, 63,000 tonnes of organic carbon, 104,000 tonnes of nitrogen and
14,800 tonnes of phosphorus were discharged to the environment after treatment process.
Since the amount of the valuable organic carbon, nitrogen and phosphorus are substantial but
not yet fully utilized, it is worth knowing what the exact fate of these valuables are in the
wastewater treatment process. Mass balance is a measure to track a mass loading of a substance
in the inlet(s) and in the outlet(s) of a chemical treatment process (Tchobanoglous et al., 2003).
It is commonly used to study the fate or the distribution pattern of a substance in the wastewater
treatment process (Shi, 2011, Wett et al., 2007, Gans et al., 2007). Commonly, a wastewater
treatment process includes mechanical primary settlement and secondary biological treatment
(Tchobanoglous et al., 2003). A complete process will also have sludge treatment associated
with anaerobic digestion (Shi, 2011, Wett et al., 2007). . Since organic carbon is commonly
14
expressed as COD in wastewater treatment, COD is used to represent organic carbon in the
following content of this chapter.
2.2.1 Fate of COD, N and P in primary settlement
2.2.1.1 Process mechanism

Primary settlement aims to remove the readily settleable solids and floating material by
sedimentation, but soluble organic content including contaminants, are not typically removed
at this stage, unless there is chemical addition. Theoretically, the settleable particulate
contaminants are removed after and end up in the primary sludge (Tchobanoglous et al., 2003).
2.2.1.2 COD, N and P Mass balance of primary settlement

COD
The typical contribution of the particulate fraction to total wastewater COD is approximately
60% (Henze and Comeau, 2008, Tchobanoglous et al., 2003). Commonly, the COD removal
rate of primary settlement ranges from 30-40 %, with the remaining 60-70% of COD flowing
to secondary treatment (Shi, 2011, Wan et al., 2016, Gans et al., 2007).
Nitrogen
Typically, the soluble ammoniacal nitrogen (NH4+-N) contributes 60-75 % to the total nitrogen
(TN) content (Tchobanoglous et al., 2003, Henze and Comeau, 2008). Since the soluble nitrites
or nitrates are only a minor component of TN, the majority of the remaining 40 % of TN content
is present as organic nitrogen (both particulate and soluble fractions) (Henze and Comeau,
2008). Therefore, theoretically no more than 40 % of TN is captured in the primary sludge. In
practice the percentage of TN captured in primary sludge is only 10-25 % (Kristensen et al.,
2004, Shi, 2011). Therefore, 75-90 % of TN in wastewater treatment works flows to the
secondary treatment process.
Phosphorus
The soluble content of phosphate (PO43--P) is the dominant component of the total phosphorus
(TP) in raw wastewater, and the typical contribution of it to the TP content is 67%
(Tchobanoglous et al., 2003, Henze and Comeau, 2008). Therefore, theoretically less than 33 %
of the inflowing TP is captured in the primary sludge. In practice, the percentage of P captured
15
in primary sludge in a pure mechanical primary settlement is only 10-25 % (Kristensen et al.,
2004, Shi, 2011). However, PO43--P can be removed via the addition of aluminum (AlCl3 or
Al2(SO4)3) or ferric salts (FeCl3 or Fe2(SO4)3) (Tchobanoglous et al., 2003). Depending on the
chemical dosage, the TP removal rate can be up to 75-95% (Tchobanoglous et al., 2003). The
studies of Kristensen et al. (2004) and Grizzetti and Bouraoui (2006) shows that, if chemical
dosing is involved in primary treatment, 20-92 % of the TP is removed and captured in the
primary sludge. Therefore, if chemical dosing is not in place, 75-90 % of the TP to the treatment
works will flow to the secondary treatment; if chemical dosing is used then 18-80% of the TP
of the work will flow to secondary treatment. There is a great variation in TP removal in
primary settlement.
2.2.2 Fate of COD, N and P in secondary settlement
Secondary biological treatment utilizes microorganisms to treat the COD, nitrogen and TP. In
all of the WWTPs investigated the activated sludge process is employed as the secondary
treatment process. This Section 2.2.2 therefore only reviews the fate of the targeted substances
(COD, N and P) in the activated sludge process.
2.2.2.1 Process mechanism of activated sludge

COD
For COD removal, aerobic heterotrophic microorganisms oxidize organic matter into CO2 and
water (H2O). These microorganisms assimilate at the same time part of the organic matter
present in the secondary treatment for its own growth (Tchobanoglous et al., 2003). Oxidation
requires that microorganisms have to be maintained at a constant level to deliver satisfactory
COD removal efficiency. However, the excess growth of microorganisms from the organic
matter assimilation are removed as Surplus Activated Sludge (SAS). Thus, after the secondary
treatment, COD, namely the organic carbon, is distributed between CO2, the secondary
treatment effluent, and the SAS (Puig et al., 2008).
Nitrogen
There are three elements to nitrogen removal (Tchobanoglous et al., 2003):
• Nitrification, in which aerobic autotrophic nitrifiers will oxidize the ammoniacal
nitrogen to nitrite (NO2-) and nitrate (NO3-)
16
• Denitrification may also occur when wastewater is recirculated to an anoxic zone.
Denitrifiers will reduce the NO2- and NO3- produced during nitrification to N2
• Assimilation of nitrogen during microorganism growth
Thus, after secondary treatment, nitrification and denitrification may occur, and total nitrogen
content may be distributed between N2, the secondary treatment effluent and the SAS. If
denitrification or nitrification do not occur, nitrogen content is distributed between the
secondary treatment effluent and the SAS.
Phosphorus
TP removal is accomplished via simultaneous assimilation during microorganism growth, and
this effect can be enhanced by installing an anaerobic zone prior to the aerobic zone
(Tchobanoglous et al., 2003). Sedlak (1991) describes the process as follow:
1. The Phosphate Accumulated Organism (PAO) in the anaerobic zone break down the
energy rich polyphosphate in their cells, for the energy needed for the production of
Polyhydroxybutyrate (PHB) through assimilation of the fermented product, such as
volatile fatty acid. Meanwhile, phosphate is released through the breakdown of
polyphosphate, and hence increasing phosphate in the wastewater.
2. In the aerobic zone, energy is produced by the oxidation of the PHB and phosphate is
stored back into the cells. Because some bacteria can store more polyphosphate than
they need in their cells, the assimilation during microorganism growth will increase
phosphate removal.
TP is only distributed between the secondary treatment effluent and the SAS (Puig et al., 2008).
2.2.2.2 COD, N and P Mass balance of activated sludge

COD
Puig et al. (2008) and Meijer et al. (2001) both experimentally measured the COD loading of
influent and effluent of the activated sludge process and SAS. Meijer et al. (2001) suggested
that the COD load in the SAS is 30-60% of the total removed COD loading. If activated sludge
is able to remove 90% of the influent COD, 27-54 % of the COD to activated sludge will be in
SAS, and 36-63 % of the COD is emitted as CO2. Puig et al. (2008) indicated that
17
approximately 42.5 % of the inflowing COD to the activated sludge ended in the SAS, and
approximately 49.5 % of the COD was emitted as CO2. Gans et al. (2007) reported a similar
pattern as 46.8% of the inflowing COD ends as CO2 and 43.5% of it is in the SAS. However,
Shi (2011) suggested that 67.6% of the COD could be emitted as CO2, whilst the SAS only
captured 23.3% of the COD. Meanwhile, a case study of the Strass WWTP in Austria suggested
that, a high rate activated sludge system is able to act as a primary treatment (Wett et al., 2007).
It can capture 60% of the inflowing COD into the SAS (Wett et al., 2007). There is a great
variation of the COD distribution pattern in the secondary biological treatment of activated
sludge.
Nitrogen
Sötemann et al. (2006) modelled the TN mass balance of nitrification of the activated sludge
process and suggested that approximately 22% of the N entering the activated sludge process
was captured in the SAS. Shi (2011) found approximately 12.3% of influent N was captured in
the SAS in an activated sludge process with denitrification, while 47.7% was released as N2
gas and 40.0% remained in the final effluent. In Gans et al. (2007), 18.2% of the influent in the
SAS was TN, 15.0% was found released as N2 gas, and 67.0% was discharged as final effluent.
There was also a great variation of the nitrogen distribution pattern in the secondary treatment.
Phosphorus
Puig et al. (2008) experimentally measured the TP loading of the influent and effluent of an
enhanced biological phosphorus removal (EBPR) activated sludge process and found
approximately 70% of the influent TP was captured in the SAS. However, this pattern is not
likely representative of the P distribution pattern for the WWTPs do not have EBPR. Gans et
al. (2007) and Shi (2011) reported a similar pattern that approximately 50% of the influent TP
to the activated sludge is captured in the sludge and the rest remains in the final effluent. Similar
to the COD and TP, the TP distribution pattern in the secondary treatment varies from site to
site.
18
2.2.3 Fate of COD, N and P in sludge treatment
Sludge thickening and dewatering

Sewage sludge is commonly treated with anaerobic digestion (AD). In 2014/2015, 80% of the
UK sewage sludge was treated with digestion (OFWAT, 2016). AD requires that feed sludge
contains an adequate moisture level (Mills et al., 2014). Sludge thickening and dewatering is
used to increase the solids content and it is achieved by mechanical filtering with the aid of
gravity and centrifugation (SNF Floerger, 2003). Because small particles and dissolved
chemicals pass through the filter, the filtrate typically retains some COD, N and P (Shi, 2011,
Tchobanoglous et al., 2003, Henze and Comeau, 2008). Commonly the filtrate produced from
sludge thickening and the centrate produce from sludge dewatering (via centrifuge), usually
referred as sludge returned liquor, is recirculated to the head of the works and mixed with raw
wastewater for treatment (Shi, 2011). The sludge return liquor can contribute from 20 to 90%
of the original COD loading to the WWTP (Jardin et al., 2006, Fattah, 2012, Evans, 2007).
Table 2.2 shows the ratio of COD, TN and TP loading in the sludge return liquor to the influent
loading found in Gans et al. (2007) and Shi (2011). The data suggests that there is also great
variation in the percentage of COD, TN and TP distributed in the sludge return liquor.
Table 2.2 The ratio of COD, TN and TP loading in the sludge return liquor to the influent loading reported in
literatures
COD TN TP
Gans et al. (2007) 38.4% 24.6% 62.5%
Shi (2011) 8.3% 11.9% 39.4%
In some cases, sludge thickening and dewatering is conducted prior to the anaerobic digestion
(Shi, 2011, Wett et al., 2007), where part of the COD, TN and TP are kept in the sludge return
liquor. Loading of COD, TN and TP to the digester shall be less than loading in the sludge
generated in the wastewater treatment.
Anaerobic digestion
In anaerobic digestion, the organic matter goes through three decomposition stages:
1. Hydrolysis which converts particulate matter to soluble compounds
2. Fermentation which degrades amino acids, sugars, and some fatty acids into acetate,
CO2 and hydrogen
19
3. Methanogenesis which converts is the acetate, CO2 and hydrogen into biogas of CO2
and methane.
Forty to 53 % of the COD from sludge was reported to be recovered and stored in the form of
methane (Shi, 2011, Wett et al., 2007). With respect to TP, its loading in the feed and digested
sludge does not change (Shi, 2011). In the case of TN, a portion may be emitted as biogas
during digestion, but usually in very small amounts as reported by Rasi (2009).
2.2.4 COD, N and P mass balance of wastewater treatment plants
Apart from the mass balance conducted by Shi (2011), other studies described in previous
sections did not present a holistic mass balance approach for COD, N and P in a total
wastewater and sludge treatment. For instance, Wett et al. (2007) presented a plant wide
(including wastewater treatment and sludge treatment with anaerobic digestion) mass balance
for only COD. Gans et al. (2007) showed a plant wide mass balance of COD, N and P, but the
investigation did not include anaerobic digestion. Therefore, more holistic mass balance studies
cover the COD, N and P are needed.
Wastewater treatment is a sequence process, the variation of material distribution pattern will
be amplified after each individual process is combined. If the mass balance is to serve the
purpose of process optimization, different recommendation will be given based on different
mass balances. Using the mass balances conducted on two WWTPs by Shi (2011) and by Wett
et al. (2007) as examples, the former only captured 48.4% of its COD inflowing to the work
into the sludge, but the latter was able to capture 74.8%. Later on, Shi (2011) applied a digester
that was not as efficient as the one applied by Wett et al. (2007). Shi (2011) only recovers 17.9%
of its COD inflowing to the work into biogas whilst Wett et al. (2007) recovers 35.4%.Then,
the dewatered digested sludge described in Shi (2011) still contains 22.3% of its COD
inflowing to the work, but the percentage in Wett et al. (2007) is 37.6%. In the case of Shi
(2011), improvement of the energy balance focused on producing more sludge from the
wastewater treatment and to elevate the energy recovery efficiency of its digestion (Shi, 2011).
For Wett et al. (2007), improvement of the energy balance focused on the energy saving of the
treatment process instead of elevating the biogas production (Wett et al., 2007).
20
2.3 Applying modelling of mass balance investigation and the impact of waste recovery
In previous sections the mass balance calculation method is used to study the fate of COD, N
and P in WWTPs. Since chemical energy recovery is obtained from the chemical organic
content, the chemical energy balance comes from COD or TS mass balance (Mills et al., 2014,
Wan et al., 2016). Mass balance is therefore a very useful technique for process evaluation,
designing and benchmarking WWTP (Yeshi et al., 2013).
For constructing the mass balance of a substance or an element (in a set time period), it is
important to know its mass loading in and out of the process. As the mass loading is calculated
by multiplying the flow and concentration (Tchobanoglous et al., 2003), both flow and
concentration measurements are then crucial to the construction and the accuracy of a mass
balance.
2.3.1 Wastewater Mass balance Constraints
The concentration data, such as COD, N and P, can be monitored regularly by onsite operators
in a WWTWs, or via a dedicated sampling campaign (Puig et al., 2008). A proper flow
measurement is sometimes difficult to conduct. Scottish Water and Northumbrian Water are
both Industrial sponsors of this Project and provided required information on the case study
sites. Scottish Water regularly monitors 313 of its wastewater treatment facilities and
Northumbrian Water regularly monitors 160. However, there is no facility has proper flow
measurement on all its wastewater or sludge streams (personal communication, M Haffey,
Scottish Water, April 2015, personal communication, L Wilkinson, Northumbrian Water, June
2015). In some cases, the subtraction method can be used to estimate the non-measured flow
or mass loading. For example, the gas product of activated sludge, the N mass loading in the
gas product of activated sludge can be calculated by subtracting the combined N loading in
the final effluent and the SAS from N loading of the activated sludge influent(Puig et al., 2008).
However, if more than one stream, then mass balance calculations will be severely hindered.
In some cases, the concentration data of many chemical parameters is also difficult to obtain
because it is difficult to collect a sample. For instance, some WWTPs discharge its primary
sludge through underground pipes, then the discharged primary sludge will instantly mix with
21
other sludge once it reaches the sludge storage tank Then, it is impossible to collect the primary
sludge sample in those WWTPs. Constructing mass balance via direct measurement of flow
and concentration sometimes is hindered by operational design and physical structure of the
WWTPs.
2.3.2 Applying modelling for mass balance construction
Modelling is a good tool for constructing the mass balance of wastewater and sludge treatment.
A model is comprised of mathematical equations (Henze et al., 2000, Takács et al., 1991). It is
used to represent the construction and operation of a system (Maria, 1997). In wastewater
modelling, data of tank sizes, flow rate, dissolved oxygen level of the biological treatment, are
used to construct the model that represent the construction of the treatment work (Langergraber
et al., 2004, Rieger et al., 2012). After the model is constructed, it requires a calibration stage
(Langergraber et al., 2004). The principle of the calibration is to provide the model the known
input, then compare the modelled output to the known output (Maria, 1997). The know input
is the measured flow rate and contaminant concentration of the raw wastewater, and the outputs
are usually the contaminant concentration of the effluent of treatment process (Liwarska-
Bizukojc et al., 2011, Fall et al., 2011). If there is a substantial difference between the known
input and the modelled output, then the parameters which are found influential to the modelled
output will be adjusted to a level which can leads to a convergence (Fall et al., 2011, Liwarska-
Bizukojc et al., 2011, Rieger et al., 2012, Sin et al., 2011). Once the model is validated, it can
be used to describe the behavior of the system and to predict the unmeasurable physical and
chemical variables, for instance flow and contaminant concentrations (Jones et al., 1989, Gasso
et al., 2002).
The wastewater model includes the Takacs model for settler modelling and Activated sludge
model No.1 (ASM1), ASM2, ASM2d, ASM3, and the Anaerobic Digestion Model No.1
(ADM1) for the biological wastewater and sludge treatment modelling (Takács et al., 1991,
Batstone et al., 2002, Henze et al., 2000). In the Takacs model, the mathematical equation is
used to express the settling velocity of the solid in ten virtual layers of settler for modelling the
solid settling (Takács et al., 1991). In the biological models, the COD, nitrogen and phosphorus
are first characterized into the state variables which can or cannot contain biological reactions.
The biological reactions which are expressed in mathematical equations consist of the
22
concentrations of the state variables and kinetics to form a matrix (Henze et al., 2000). Solving
the matrix will predict the COD, nitrogen and phosphorus concentrations of the treated effluent
(Fall et al., 2011, Petersen et al., 2002, Liwarska-Bizukojc et al., 2011). Moreover, inorganic
precipitation, such as struvite, can be also modelled (Musvoto et al., 2000, Hydromantis, 2017).
Hence, wastewater modelling can be applied for wide mass balance studies. For instance, the
COD mass balance conducted in Strass WWTP was performed by modelling (Wett et al., 2007).
Puchongkawarin et al. (2015) also used modelling to predict the mass balances of COD,
nitrogen derivatives, and phosphorus derives under different operational strategies.
Wastewater modelling is usually conducted by dedicated commercial software (Gernaey et al.,

2004). Commonly used wastewater modelling software includes Aquasim developed by Eawag
(Tas et al., 2009), BioWin developed by EnviroSim (Liwarska-Bizukojc et al., 2011), GPS-X
developed by Hydromantis (Fall et al., 2011), SIMBA developed by inCTRL Solutions (Wett
et al., 2007), STOAT developed by Water Research Centre (Sarkar et al., 2010), and WEST
developed by Mike powered by DHI (Vanhooren et al., 2003) as some examples.
2.3.3 Applying modelling for studying waste recovery
Mass balance of WWTPs is required to understand the fate of valuable substances including
energy during the treatment process. More importantly, it could aid to identify the potential
waste recovery opportunities, namely where and how to recover the valuables. Sometimes,
implementation of new processes or technologies may solve one problem but raise new ones
at the same time. Yeshi et al. (2013) reported that after applying modelling techniques,
predicted results such as capturing more COD in the primary sludge lead to a better chemical
energy recovery but weaken denitrification in the subsequent biological process since less
carbon source is available to the denitrification. Anaerobic treatment may bring a better energy
balance to the wastewater treatment work but it is usually incapable of removing nutrients and
hence post treatment polishing is required (Chernicharo, 2006).
Wastewater modelling provides convenience on studying the impact caused by the

implementation of a new process. This is because, reconfiguration and experimentation are
impractical and too costly to do on a real system but can be done theoretically on a model with
low budget (Maria, 1997).
23
Wastewater modelling has been used for studying the impact of implementing waste recovery
technologies. Drewnowski et al. (2018) applies modelling to predict the increase of biogas
production over the increase of suspended solid removal efficiency of the primary settlement.
Sutton et al. (2011) modelled a better the energy balance due to the virtual implementation of
anaerobic digestion for sludge treatment. Puchongkawarin et al. (2015) designed a WWTP that
consists of recovery technologies of up-flow anaerobic sludge blanket for energy recovery and
ion exchange associated with struvite crystallizer for nutrient recovery. It then used modelling
to determine mass and energy balance under different operational configurations and
discovered the best one with the greatest cost benefits. Fabiyi et al. (2016) modelled the
WASSTRIP process which aims to release the ortho-phosphate stored in the SAS produced
from the EBPR process. The modelling robustly predicted the ortho-phosphate release in the
sludge handling process and provided insight for phosphate recovery for fertilizers.
2.4 Research Questions
According to Section 2.1, although COD is well used as a surrogate of wastewater chemical
energy, the relationship between the two is poorly defined due to the time-consuming drying
process hinders the large amount of experimental measurements. Moreover, the relationship
between the TS and VS content and the sludge chemical energy needs to be validated. On this
basis, the research questions proposed are:
I. Is there a reliable and practical method to dry the wastewater but without losing a
substantial amount of energetic COD?
II. What is the best surrogate(s) of the chemical energy content of wastewater?
III. What is the relationship between the surrogates and chemical energy content of
wastewater and sludge four investigated WWTPs of the sponsor companies?
According to Section 2.2, although the fate of the COD, TN and TP has been studied based on
theory and in real treatment works, there may be a substantial variation of their distribution
patterns in each individual treatment process. Furthermore, there are few studies that perform
plant wide holistic mass balance investigation of COD, N and P together. The related research
questions proposed is:
24
I. What are the chemical energy balance and COD, TN and TP mass balances of the four
investigated WWTPs of the sponsor companies?
Understanding the fate of chemical energy, COD, N and P is to discover the potential
opportunities to improve value recovery. However, according to Section 2.3, the
implementation of new processes or technologies may cause issues to the treatment
performance. Therefore, the related research questions are:
I. From the chemical energy balance and mass balance studies, what are the opportunities
for improving the current value recovery?
If the measures for improving value recovery are implemented, what are the impact on the
current chemical energy and mass balance and on the value recovery?
25
Chapter 3 Methodology
This chapter describes the sample collection (Section 3.1), the experimental procedures for
chemical analysis (Section 3.2), the statistical analysis (Section 3.3), the procedures used for
wastewater and sludge modelling (Section 3.4), and the construction of mass and chemical
energy balances for the wastewater treatment plants (Section 3.5).
3.1 Sample and data collection methods
3.1.1 Wastewater and sludge sampling
3.1.1.1 24-hour composite sample collection

Four wastewater treatment plants (WWTP) were sampled. Each WWTP wastewater has
mechanical primary settlement and activated sludge as secondary biological treatment. Apart
from WWTP A’s Trickling Filter process, each wastewater and sludge stream was sampled,
but with varied sampling methods.
Composite sampling was applied to the collection of:

1. Raw inlet wastewater, after it had been screened and de-gritted
2. Effluent from primary treatment
3. Effluent from secondary biological treatment
Samples were collected hourly using an Aquacell (Aquamatic, UK) autosampler over a period
of 24 hours from the open channel flows in or out of the process. Hourly samples of 400 mL
were collected in 24 clean plastic 800 mL containers in the autosampler. 200 mL of sample
from each container was taken and combined to form a 24-hour composite sample of 4.8 L
volume. The compositing was time proportional rather than flow proportional.
The sludge is essentially a liquid since commonly more than 50% of its mass is water content
(Tchobanoglous et al., 2003). It is therefore collectable in the same way as a wastewater sample.
Sludge and sludge return liquor (the rejected wastewater generated from sludge thickening and
dewatering) was collected by grab sampling as they are pumped through underground piping
which is not accessible to the autosampler.
26
In an effort to sample the same wastewater during its passage through the treatment plant (i.e.
to take account of hydraulic residence time (HRT)), sampling of the raw wastewater, primary
effluent and secondary effluent was staggered according to the estimated HRT of each unit
process.
Both WWTP C and D have onsite sludge thickening and dewatering. The removed water
(called filtrate or centrate in this thesis) was recirculated to the head of the works for treatment.
The recirculation is through underground pipe and goes into the primary settlement separately
from the raw inlet. Therefore, the filtrate or centrate was sampled separately. Using WWTP C
as an example, if the sample period of the raw inlet is from 08:00 day 1 to 08:00 day 2, the
sludge return liquor of primary sludge thickening is sampled by grab sampling during the
period of 08:00 day 1 to 08:00 day 2.
3.1.1.2 24-hour spot sample collection

In order to understand the diurnal variation of flow and chemical composition of raw
wastewater hourly samples were collected over 24 hours using an Aquacell autosampler
(Aquamatic, UK). Unlike the composite sampling, each hourly sample had a volume of 800
mL and was retained in separate sample bottles.
3.1.2 Sample storage
The grab wastewater samples and the sample of the 24-hour study were stored in 1 L
Polyethylene Terephthalate (PET) rectangular sample bottles (Medfor, UK). The grab sludge
sample was stored in 500 mL Polypropylene (PP) sample jars.
Samples were stored in a cool box with ice packs during transportation. The transportation time
was from 30 minutes to 4 hours. All sludge and wastewater samples were stored in a fridge at
4 oC immediately upon arrival back at the laboratory.
Chemical analysis was typically expected to be done within 48 hours after sample collection.
If not, samples were frozen at -20oC to inhibit microbial activity which affects the chemical
composition of the sample.
27
3.1.3 Flow and energy consumption data collection
3.1.3.1 Flow data collection

Flow data were downloaded from water companies’ database after sampling was completed.
Downloaded flow data were either a series of instantaneous flow rate values recorded every 15
minutes or daily total flow from 00:00:00 to 23:59:59.
Since the sampling period for composite samples was 24 hours the 24-hour total flow was also
appropriate for the mass and energy balance study. For treatment plants that recorded flow at
15-minute intervals, these measurements were summed to provide a 24 hour total.
Because most of the composite sampling did not start at exactly 00:00:00 and typically spanned
two days, if only daily total flow was available, the daily total flow of those two consecutive
days was collected. 24 hour flow for the period of sampling was then calculated proportionately
according to the individual flows on the two days e.g. if sampling was undertaken for 14 hours
during a day when the flow was 10 000 m3/d and 10 hours when flow was 8 000 m3/d then the
daily flow was calculated as: [10 000 × (14÷24)]+[8 000 × (10÷24)].This calculation is based
on an assumption that the hourly flow is the same during a day. The flow of WWTP D is
calculated through this approach.
3.1.3.2 Energy consumption data collection

All the energy data were downloaded from water companies’ databases after the sampling was
finished. Original energy data downloaded were either a series of half-hourly consumption or
daily total which counts energy consumption from 00:00:00 to 23:59:59 of a day.
For data recorded as half-hourly consumption, the 24-hour consumption was then obtained by
summing the half-hourly consumption during the sampling period.
If only daily total energy consumption was available, based on the assumption that the energy
consumption rate stays constant during a day, the 24-hour consumption was then calculated
proportionately according to the individual consumption on the two days, e.g. if sampling was
undertaken for 14 hours during a day when the electricity consumption was 6,000 kWh/d and
10 hours when flow was 7,200 kWh/d, then the daily electricity consumption was calculated
28
as [6,000 × (14÷24)]+[7200 × (10÷24)]. The electricity consumption of WWTP A, C and D
are calculated through this approach.
3.2 Chemical analysis
3.2.1 Chemical analysis of wastewater
Chemical analysis of wastewater included Chemical Oxygen Demand (COD), soluble COD
(sCOD), particulate COD (pCOD), Total Organic Carbon (TOC), Dissolved Organic Carbon
(DOC), particulate Organic Carbon (POC), Ammoniacal Nitrogen (NH4+-N), Total Kjeldahl
Nitrogen (TKN), Nitrite Nitrogen (NO2--N), Nitrate Nitrogen (NO3--N), Total Nitrogen (TN),
Phosphorus as Phosphate (PO43--P), Total Solid (TS) by oven drying, Volatile Solid (VS), Total
Solid (TScve) by centrifugal vacuum evaporation, and chemical energy content.
Composite samples collected in the mass and energy balance sampling were analysed for all
the variables. The grab samples of sludge return liquor were also analysed for all variables
except for TScve and chemical energy content.
Samples collected in the 24-hour study were analysed for COD, sCOD, pCOD, NH4+-N and
PO43--P. Starting with the 1st hourly sample, in total 6 samples were selected at 4-hourly
intervals and measured for TScve and chemical energy content.
NO2--N and NO3--N were measured in duplicate, TScve was measured in either duplicate or
triplicate, chemical energy was measured once on each sample, and all other parameters were
measured in triplicate.
3.2.1.1 Sample preparation

Samples for sCOD, DOC, NH4+-N and PO43--P were filtered through 0.45 µm Polyether
Sulfone (PES) syringe filters (VWR, US). Samples for NO2--N and NO3--N were filtered
through 0.20 µm Polyether Sulfone (PES) syringe filters (VWR, US). There was no sample
preparation for other variables.
29
3.2.1.2 COD measurement
COD measurement required 0.3-3.0 mL of sample and was conducted with a Spectroquant ○
R
COD Cell Test (Merck Millipore, Germany) with measuring ranges of 4-40 mg/L, 5-80 mg/L
and 25-1500 mg/L and Spectroquant ○
R Phare 300 Spectrometer (Merck Millipore, Germany).
For sCOD measurement, 0.45 µm filtered sample was analysed following the COD
measurement procedure. pCOD is calculated by subtracting sCOD from COD.
3.2.1.3 Organic carbon measurement

TOC measurement required 0.1-2.0 mL of sample and was conducted with a TOC Cuvette Test
(Purging Method) (Hach, US) with measuring range of 3 – 30 mg/L and 30-300 mg/L and
DR1900 Portable Spectrometer (Hach, US).
For DOC measurement, 0.45 µm filtered samples were analysed following the TOC
measurement procedure. POC is calculated by subtracting the DOC from TOC.
3.2.1.4 Nitrogen measurement

NH4+-N measurement required 0.2-1.0 mL of 0.45 µm filtered sample and was conducted with
a Spectroquant ○
R Ammonium Cell Test (Merck Millipore, Germany) with a measuring range
of 0.2-8.0 mg/L and Spectroquant ○

R Ammonium Test (Merck Millipore, Germany) with a
measuring range of 2-75 mg/L. NH4+-N was analysed with a Spectroquant ○

R Phare 300
Spectrometer (Merck Millipore, Germany)
TKN measurement was adapted from standard method 4500-NOrg C (APHA, 2012). For this
measurement 0.1–10 mL of sample was required. It was mixed in a Kjeldahl Flask with 14 mL
of concentrated H2SO4 and 2 tablets each containing 5 g of K2SO4 and 0.5 g CuSO4∙5H2O
(Gerhardt, Germany). The solution was digested at 400 oC for 60 minutes with a Turbotherm
heater (Gerhardt, Germany) to convert TKN into ammonium. After the digested solution was
cooled to ambient temperature it was mixed with 80 mL 40% NaOH solution and distilled in
Vapodest 30s (Gerhardt, Germany) to convert NH4+ into free NH3 and transferred to a flask
containing 50 mL of a 20 g/L H3BO3 solution. The H3BO3 solution was then titrated with 0.05
N H2SO4 solution to determine the TKN concentration.
30
NO2--N and NO3--N measurement required 5-6 mL of 0.20 µm filtered sample and were
analsyed using an Ion Chromatograph (IC) Dionex ICS-1000 with Ionpack AS 14A as column
and carbonate as eluent.
TN is the sum of TKN, NO2--N and NO3--N.
3.2.1.5 Orthophosphate measurement

PO43--P measurement requires 0.2-1.0 mL of 0.45 µm filtered sample and was conducted with
Spectroquant ○
R Phosphate Cell Test (Merck Millipore, Germany) with measuring ranges of
0.05-5.0 mg/L, 0.5-25.0 mg/L and 3.0 – 100.0 mg/L and Spectroquant ○
R Phare 300
Spectrometer (Merck Millipore, Germany).
3.2.1.6 Solids measurement

TS and VS measurement used 40 mL of sample and follows standard method 2540 B and 2540
E (APHA, 2012).
For the TScve measuremernt the sample flask (Genevac, UK) was pre-dried at 103-105 oC until
its weight change was less than 4% which is the allowable variation between two TS
measurement in standard method 2540 B (APHA, 2012). The weight of the pre-dried container
was recorded as mf. Sample with a volume (v) of 300-400 mL was placed into the pre-dried
container and dried by centrifugal vacuum evaporation using a Rocket 4D Synergy (Genevac,
UK) at 18 mbar, 30oC and 1,800 rpm for 19 hours and 40 minutes to remove all the visible
water. After the centrifugal evaporator drying, the container (with dried residue in it) was
further dried in a dessicator at ambient temperature for at least two days until its weight change
was less than 4%. The weight of the container was recorded as mf+dr. TScve was calculated by
Equation 3-1.
Equation 3-1
(𝑚𝑓+𝑑𝑟 − m𝑓 ) × 1000
𝑇𝑆𝑐𝑣𝑒 =
𝑉
Where mf+dr is the weight of the container with dried residue in it, mf is the weight of the
container, V is the volume of the sample placed into the container.
31
3.2.1.7 Chemical energy measurement of wastewater
The dried residue produced from the TScve measurement was used for analysis of the chemical
energy content of the same aliquot.
A Parr 6100 Compensated Jacket Calorimeter (Parr Instrument Company, USA) was used for
energy measurement. The bomb calorimeter requires approximately 26 kJ of heat in one
measurement but no more than 2 g of sample input. Thus, the calorific value of the sample in
the calorimeter has to be higher than 13 kJ/g. Previous reports of the calorific value of dried
wastewater are in the range 3.2 to 10.5 kJ/g (Shizas and Bagley, 2004, Heidrich et al., 2010),
and therefore the dried residue needed to be co-combusted with a high calorific value substance.
Paraffin wax, which has been used as the combustion aid in the combustion study (via bomb
calorimetry) of Germanium metal (Sunner and Månsson, 1979), was chosen as the combustion
aid in this study. Its calorific value was determined as a mean value of 46.7344 kJ/g via
experimental bomb calorimetry on 8 samples with weight 0.6001–0.6409 g.
0.15 – 0.60 g of dried wastewater sample was then mixed with 0.40 – 0.60 g of Paraffin wax
to form approximately 1 g of combustible sample to be introduced to the bomb calorimeter.
The weight of the dried sample, and the weight and calorific value of the paraffin wax, were
entered into the instrument, and the calorimeter then automatically calculated the dried
sample’s calorific value (kJ/g) after each measurement. The chemical energy content of
wastewater (Eww) was calculated by multiplying the calorific value of the dried wastewater
(Edww in kJ/g) of the dried sample with TScve (in mg/L) of the sample, as shown in Equation 3-
2.
Equation 3-2
𝐸𝑤𝑤 = 𝐸𝑑𝑤𝑤 × 𝑇𝑆𝑐𝑣𝑒
Where 𝑬𝒅𝒘𝒘 is the caloric value of the dried residue, TScve is the TS concentration measured
by the centrifugal vacuum。
40 mL of the same wastewater sample used for the TScve and chemical energy measurement
was also separately dried by the Rocket 4D Synergy under the same conditions. After drying,
40 mL of deionised water was added into the flask for rehydration. The samples underwent
32
sonication for 10 minutes to ensure complete rehydration. COD concentration of the rehydrated
sample was then measured in triplicate. The COD recovery rate was then calculated by dividing
the original COD by the rehydrated COD. Multiplied by 100 this gave the percentage of COD
retained in the dried sample. Calculation of COD recovery rate is shown in Equation 3-3
Equation 3-3
𝐶𝑂𝐷(𝑜)
𝐶𝑂𝐷 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑦 𝑅𝑎𝑡𝑒 =
𝐶𝑂𝐷(𝑟𝑒ℎ𝑦)
Where COD(o) is the COD concentration of the original wastewater before drying, and COD(rehy)
the COD concentration of the rehydrated wastewater.
The energy content so determined (Eww) was that used to explore the relationship with COD
and to calculate the energy balance of the WWTPs.
3.2.2 Chemical analysis of sludge
All the sludge samples collected were analysed for Total Solid percentage (TS%), Volatile
Solid percentage to TS (VS/TS%), Total Organic Carbon percentage to TS (TOC/TS%), Total
Nitrogen percentage to TS (TN/TS%), Total Phosphorus percentage to TS (TP/TS%) and
chemical energy content. However, in some specific cases variables were not measured due to
insufficient sample size. These individual cases are highlighted in the Results & Discussion
chapters.
3.2.2.1 Sample preparation

Apart from the TS% and VS% measurement and TOC% measurement of the SAS, sludge was
dried in an oven at 103-105 oC until constant weight was reached prior to any further
measurements.
For TOC%, TN% and TP%, the dried sample was ground to a fine powder using a James Martin
ZX809X Spice and Coffee Grinder (Wahl Clipper Corporation, USA) before measurement.
3.2.2.2 Solid measurement

TS% and VS/TS% were measured in the same set of experiments. Triplicate analyses of both
variables were undertaken on each sludge sample.
33
The method of TS% and VS/TS% measurement is adapted from standard method 2450 G
(APHA, 2012). The adaptation is that the sample mass is reduced from the recommended 25-
50 g to 3-5 g due to limited oven and muffle furnace space. In terms of procedures, a crucible
was dried in a muffle furnace at 550oC for 1 hour and then stored and cooled in a desiccator
overnight. After the weight (mc) of the crucible was recorded, 2.0 to 5.0 g of sludge sample
was added to the crucible. The crucible was weighed again to obtain mc+s. The crucible was
then dried in the oven at 103-105 oC overnight. After drying, the crucible was cooled in a
dessiccator until its weight change was less than 4%. The weight of the crucible was recorded
as m104. TS% was then calculated using Equation 3-4.
Equation 3-4
𝑚 −𝑚
TS% = 𝑚104 −𝑚𝑐 × 100%
𝑐+𝑠 𝑐
Where m104 is the weight of the crucible with dried residue after oven drying, mc+s is the weight
of the crucible with wet sludge sample, mc is the weight of the crucible after 550oC muffle
furnace drying.
The crucible was then dried in a muffle furnace at 550oC for at least 30 minutes. After drying,
the crucible was cooled in a dessiccator until its weight change was less than 4%. The weight
of the crucible was recorded as m550. VS/TS% was the calculated by using Equation 3-5.
Equation 3-5
𝑚550 − 𝑚𝑐
VS/TS% = (1 − ) × 100%
𝑚104 − 𝑚𝑐
Where m550 is the weight of the crucible with dried residue after 550oC muffle furnace drying,
m104 is the weight of the crucible with dried residue after oven drying, and mc is the weight of
the crucible after 550oC muffle furnace drying.
3.2.2.3 TOC/TS% measurement

TOC/TS% was determined using 0.1 g of 103-105 oC oven dried and ground sample. 1 mL of
4.0 mol/L HCl was added to each crucible. The HCl was then removed from the sample after4
hours. The sample was then oven dried at 60 to 70 °C for 16 – 24 hours. A Leco CS230
Carbon/Sulphur Analyser was used to measure the carbon percentage of the sample and the
blank. The TOC/TS% was calculated using Equation 3-6.
34
Equation 3-6
𝑇𝐶𝑠 % − 𝑇𝐶𝐵 %
TOC/TS% =
𝐷
Where TCs% is the measured carbon percentage of the sample, TCB% is the measured carbon
percentage of the blank, and D is the dry matter factor.
The solid content in the SAS was lower than 1% (equivalent to 1,000 mg/L). For generating
0.1 g of dried solid for the TOC% measurement, it was therefore necessary to use more than
100 mL of sample. Because of restrictions on oven drying space in the laboratory the TOC
content of SAS was measured following the experimental procedure shown in Section 3.2.1.3
and is in units of mg/L. TOC/TS% of the SAS is then calculated via Equation 3-7.
Equation 3-7
𝑇𝑂𝐶
TOC/TS% =
𝑇𝑆% ÷ 100% × 1000 × 1000
Where TOC (mg/L) is the total organic carbon of the SAS and TS% is the total solid content
of the SAS.
3.2.2.4 TN/TS% measurement

TN/TS% was measured by the Scientific Service of Scottish Water (SSSW). Samples were
dried and ground before posting. A single measurement was done on each sample.
0.05 g of sample was microwave-digested with 25 ml of Na2S2O8 or NaOH solution. The

digestion was conducted at 140oC for 40 minutes. The digested sample was then diluted to 100
mL with deionised water, and subsequently analysed with standard method 4500-N B In-Line
UV/Persulfate Digestion and Oxidation with Flow Injection Analysis (APHA, 2012). The
method firstly breaks down the simple nitrogen compound (i.e. Ammonia) and the complex
nitrogen compound to Nitrate. All the Nitrate is then reduced to Nitrite which is quantified for
determining the TN/TS% of the sample.
Although the analytical method measures the TN/TS% of the sample, NH4+-N may be lost due
to evaporation during preliminary drying at 103-105oC. Hence, the measured TN% may be
lower than the real value. Assuming only NH4+-N is lost during the drying, actual mass of TN
35
of sludge is the sum of the mass of TN in the dried solid and the mass of NH4+-N lost. Assuming
the NH4+-N is only within the liquid fraction of the sludge and its concentration is CAmmN , the
adjusted TN% of sludge is calculated by Equation 3-8.
Equation 3-8
𝐴𝑚𝑚𝑁
𝑇𝑁% × 𝑇𝑆% + 𝐶 × (1 − 𝑇𝑆%) × 𝜌𝑆𝑙𝑢𝑑𝑔𝑒
𝑇𝑁%𝑎𝑑𝑗𝑢𝑠𝑡𝑒𝑑 =
𝑇𝑆%
where TN/TS% is the measured TN content in the dried solid of sludge, TS% is the total solid
percentage of sludge, CAmmN is NH4+-N concentration of the corresponding sludge return liquor,
and the ρSludge is the density of the sludge.
As all other component of Equation 3-8 were measured, the key to the calculation is the NH4+-
N concentration in the water content of sludge. In sludge thickening and dewatering, water
fraction is removed and the NH4+-N of it is measured. By assuming the NH4+-N of the water
content of the wet and thickened/dewatered sludge are the identical to the NH4+-N of the
removed water fraction, TN%adjusted of the sludge before and after thickening or dewatering can
be obtained.
However, only WWTP C has sludge thickening and dewatering on each of the primary sludge,
SAS, sludge fed to the digester and the digested sludge. As sampling and analysis of the
thickening and dewatering facilities was undertaken on multiple occasions the
TN%:TN%adjusted of each type of sludge was determined and an average value calculated. The
TN%adjusted of sludge produced in other WWTPs was adjusted by multiplying its TN% with the
TN%:TN%adjusted of a corresponding type of sludge：the TN%:TN%adjusted of the primary
sludge calculated from the primary sludge of WWTP C will be applied to the primary sludge
produced in other three WWTPs; the the TN%:TN%adjusted of the SAS calculated from the SAS
of WWTP C will be applied to the SAS produced in WWTP A and D. The result of the
TN%:TN%adjusted and the adjusted TN% of sludges are presented in Appendix A.
3.2.2.5 TP/TS% measurement

TP% was also measured by SSSW. Samples were dried and ground before posting. A single
measurement was done on each sample.
36
0.5g of sample was mixed with 4 mL of concentrated HNO3 and 4 mL of concentrated HCl.
The solution was pre-digested in the uncapped vessel for at least 30 minutes or until any visible
reaction had ceased. The pre-digested solution was then microwave-digested in a capped vessel
at 175 oC for 15 minutes. The digested solution was filtered through a 540 filter paper
(Whatman) and then diluted to 100 mL with deionised water. The phosphorus content was then
measured using a Perkin Elmer Nexion 300X ICPMS Spectrometer at 30.994 AMU with KED
(Collision) Helium Gas flow of 5.4 mL/min and is corrected for using Scandium as an internal
standard measured at 45 AMU.
3.2.2.6 Chemical energy measurement

Chemical energy measurement of sludge followed the experimental procedure described in
Section 3.2.1.7. 0.4-0.7 g of dried sludge sample was mixed with 0.4-0.6 g of Parafin Wax to
form approximately 1.0 - 1.2 g of combustible sample which was measured for its calorific
value by the same bomb calorimeter described in Section 3.2.1.7. The chemical energy content
of dried sludge (Eds) is expressed as KJ/g. The chemical energy of the wet sludge (Es) is
calculated via Equation 3-9.
Equation 3-9
𝐸𝑠 = 𝜌𝑠 × 𝑇𝑆% × 𝐸𝑑𝑠
Where ρs is the density of the sludge and is taken as 1 g/cm3, TS% is the total solid content of
the sludge, and Eds is the energy content of the dried sludge.
3.3 Statistical analysis
Descriptive statistics, such as mean value and standard deviation was conducted using Minitab
2018.
For the study of relationships between energy content and chemical constituents of wastewater
or sludge the Correlation and Best Subset functions of Minitab 2018 were used to determine
which parameter(s) best reflect the wastewater or sludge energy content via single or multiple
regression. The graphs of regressions were produced in Minitab 2018.
37
3.4 Wastewater and sludge treatment modelling
Constructing correct mass and energy balance requires measurement of both flow and
concentration on every inflow and outflow stream of the wastewater and sludge treatment
process. However, such comprehensive sampling is not undertaken by the water companies.
Thus, existing available data is used here, first to predict the unknown flows and concentrations,
then to construct models of the four wastewater and sludge treatments for, and ultimately
construct mass and energy balances for each of them.
There are three steps to the wastewater and sludge treatment modelling in this thesis:
1. model construction (in Section 3.4.1)
2. influent characterization (in Section 3.4.2)
3. model calibration and validation (in Section 3.4.3)
In the following sections the experimentally measured values (of flow and concentration) are
referred to as ‘observed values’, and the modelled, simulated values are referred to as ‘predicted
values’.
3.4.1 Model construction
3.4.1.1 Operating model/library

The modelling work in this research was undertaken in software GPS-X 7.0. All the wastewater
and sludge treatment modelling were conducted under library “Comprehensive Carbon,
Nitrogen, Phosphorus (mantis2lib)”. All the biological processes are modelled with “mantis2”
which is an in-house model developed by the software company Hydromantis and is developed
from biological models of Activated Sludge Model 2d (ASM2d) and Anaerobic Digestion
Model 1 (ADM1) (Hydromantis, 2017). The “mantis2” is able to model the COD, nitrogen and
phosphorus mass balance of the wastewater and sludge treatment (Fall et al., 2011,
Drewnowski et al., 2018). Moreover, it can also model the application of advanced technologies
such as anaerobic wastewater treatment and Struvite recovery treatment which are not included
in the classic ASM2d model or ADM model (Hydromantis, 2017).
38
3.4.1.2 Modelling of individual treatment processes
3.4.1.2.1 Chemical dosing
The chemical dosing in the modelling is for two purposes. One is for mimicking the actual
phosphate removal by intentional chemical dosing in the real world. The other one is for
virtually removing sCOD and phosphate for the modelling purpose.
Actual Ferric dosing for phosphate removal

WWTP A does have an actual chemical dosing process for phosphate removal. Therefore, its
model uses the “In-line Chemical Dosing” unit for mimicking the process. Ferric sulphate
(Fe2(SO4)3·9H2O) is chosen as the chemical. The “percent hydrated ferric sulphate” is set as
100%. The dosage method is mass based, and the dosage rate is 236 kg/d according to the
historical data (personal communication, J McCowan, Scottish Water, August 2018).
Virtual aluminum dosing for sCOD removal

It is possible that there is a considerable reduction of sCOD after the primary treatment because
part of the sCOD first flocculates into pCOD and subsequently settles (Gernaey et al., 2001).
For mimicking this process, the “In-line Chemical Dosing” unit uses aluminum sulphate
(Al2(SO4)3·18H2O) to coagulate and flocculate the soluble inert COD and the soluble colloidal
COD into pCOD in its default settings. This research uses this unit for the purpose of virtual
sCOD removal. The dosage aimed for is the minimum required to remove sCOD. The
kineticrelated to the sCOD removal are obtained via the “Optimization” function that aims to
remove all the soluble inert COD (21.5 mg/L) and the soluble colloidal COD (39.99 mg/L)
from a wastewater with a flowrate of 10,000 m3/d, 430 mg/L COD and 147.5 mg/L sCOD via
1 kg/d dosage of Al2(SO4)3.18H2O (personal communication, S Snowling, Hydromantis,
November 2018). The kinetics of the process of sCOD removal obtained will then be applied
in the virtual sCOD removal unit if the sCOD reduction is greater than 15 mg/L after primary
settlement. The default and adjusted value of related kinetic of the process of sCOD removal
via the Al2(SO4)3·18H2O dosing are shown in Table 3-1.
39
Table 3-1 The default and adjusted values of related kinetic of the Al2(SO4)3·18H2O for virtual sCOD
removal
Default Adjusted
Value Value
pKsp, apparent - Log of soluility product of AlPO4 13.6 0.00000000
Minimum metal to colloidal organic ratio (at high concentration), g Metal/g
4.0 0.00012632
COD removal
Minimum metal to colloidal organic ratio (at high concentration), g Metal/g
20.0 33.71681100
COD removal
Affinity factor for colloidal organic, m3/g removal 0.5 1.19473060
Minimum metal to snd ratio (at high concentration), g Metal/g COD
4.0 0.00000000
removal
Minimum metal to snd ratio (at high concentration), g Metal/g COD
20.0 0.00000000
removal
Affinity factor for snd, m3/g removal 0.5 0.00000000
Minimum metal to si (at high concentration), g Metal/g COD removal 4.0 0.00020201
Minimum metal to si (at high concentration), g Metal/g COD removal 20.0 34.96696400
Affinity factor for si, m3/g removal 0.5 1.37120470
As the unit serves a virtual sCOD removal purpose, the chemical shall not react with other
possible reactive chemicals, namely the phosphate. Therefore, the ksp value of the AlPO4 is set
as “0” to prevent any phosphate reacting with the Al2(SO4)3·18H2O.
Virtual ferric dosing for phosphate removal

It is also possible that there is a considerable reduction of phosphate after the primary treatment.
This is because some phosphate may react with unexpected iron content in the primary settlers,
which scavenges phosphate and results in settling of iron phosphate (personal communication,
L Wilkinson, Northumbrian Water, October 2018). If phosphate reduction is greater than 1.5
mg/L after primary settlement, the “In-line Chemical Dosing” unit, using Fe2(SO4)3·9H2O, will
be applied. The “percent hydrated ferric sulphate” is set as 100%. The dosing method is mass
based.
3.4.1.2.2 Primary settlement
The current library has three types of primary settlers, the “Circular Primary Settlers”, the
“Rectangular Primary Settlers”, and the “High Rate Treatment”. The appropriate unit, and its
dimensions, is chosen according to the configuration of the actual WWTPs.
40
The Simple1D model that developed from Takács model is used in all the models (Takács et
al., 1991). The primary de-sludging process can be initiated and controlled through a variety
of triggers, such as a given discharge rate, a rate that is proportional to a specific discharge rate
(i.e. to the inflowing wastewater), or a proportional–integral–derivative (PID) controller which
is able to control the COD of the discharged primary sludge. The WWTP A and D are controlled
via the PID controller using the COD (estimated from the VS content via Equation 4-4,
explanation given in Section 4.2.3) of the primary sludge as the control variable. WWTP B is
controlled via a TSS concentration because the primary settlement unit used is unable to be
controlled by the COD concentration. WWTP C is controlled via a given discharge rate
according to the historical data because the primary sludge was not measured during the
sampling campaign.
Considering WWTPs A and D have surplus activated sludge recirculation, their primary settlers
are converted into a reactive settler whose biological reactions are controlled by an independent
“Plug Flow Tank” unit. The primary settlers of WWTP B and C are pure mechanical settlers.
In reality, the majority of WWTPs have more than one settler. In order to simplify the modelling,
only one primary settler is used in each modelling exercise. Its centre depth and the side wall
depth are the same as the actual ones and inputted into the model. The surface area inputted to
the model is the sum of the surface area of all the actual settlers (Hydromantis, 2017). The inlet
point of the wastewater is set equal to the water level of the tanks. The configuration of the
primary settlement tanks of each WWTP are shown in Table 3-2.
41
Table 3-2 The configuration of the primary settlement tanks of each WWTP
WWTP A WWTP B WWTP C WWTP D
Tank Circular sloping High rate treatment Rectangular sloping Circular settler
Dimension bottom settler bottom settler Centre depth - 3.75m
Centre depth - 3.6m Water depth - 3.3m Sidewall depth - 1.8m
Sidewall depth - 2.0m Surface Area - Surface Area -

9750m2 6,945m2
Surface Area -
1,029m2
Control Target COD of Target TSS Set flowrate of Target COD of

Method primary Sludge – concentration of primary sludge - primary Sludge –
34,844 mg/L primary sludge – 1,548 m3/d 24,225 mg/L
30,000 mg/L (the
Sampling interval - predicted COD of Sampling interval -
15 minutes 43,120 mg/L which is 15 minutes
close the target COD
Proportional gain - of 43,951 mg/L) Proportional gain -
0.001 0.001
Integral time - 0.01 Integral time - 0.01
3.4.1.2.2 Secondary biological treatment
The main biological treatment process used in all four WWTPs is the aerobic suspended
activated sludge process (consisting of an aerobic zone only). Therefore the “Plug Flow Tank”
unit and the subsequent secondary clarifier unit are chosen to model the biological treatment
process. The dimension of the “Plug Flow Tank”, dissolved oxygen (DO) concentration, mixed
liquor suspended solid (MLSS) concentration, and the volume or the ratio of the return
activated sludge (RAS) are set according to the actual configuration and/or the historical data
of each WWTP. The discharge of SAS is controlled by a PID controller using the measured
MLSS of the mixed liquor in the activated process as the control variable.
In WWTP A and D the secondary settlers are biologically active and linked to an independent
“Plug Flow Tank” unit which controls the biological reactions. Similar to the primary settlers,
there is only one secondary settler used in each modelling exercise. The center depth and the
side wall depth of the secondary settlers are set to be the same as the actual ones, and the surface
42
area of each is the sum of the surface area of all actual settlers. The inlet point of the wastewater
is set 2/3 of the center depth from the bottom according to recommendations provided by the
engineer from the Hydromantis (personal communication, S Snowling, Hydromantis, August
2018). The configuration of the activated sludge process and the subsequent clarifier for each
WWTP is shown in Table 3-3.
Table 3-3 The configuration of the activated sludge process and the subsequent clarification process for each
WWTP
Tank Volume - 1,924 m3 Volume – 22,291 m3 Volume - 33,203 m3 Volume - 21136.5m3
Dimension
of Activated Depth – 3.93 m Depth – 8.6 m Depth – 6.5 m Depth – 4.5 m
Sludge
DO 2.0 mg/L 2.0 mg/L 2.5 mg/L 0.66 mg/L
RAS 71% to the Primary 64,694 m3/d 225,152 m3/d 87,452 m3/d
Control Effluent
Control Target MLSS- 2,405 Target MLSS- 2,540 Target MLSS- 2,805 Target MLSS-1,733
Method mg/L mg/L mg/L mg/L
Sampling interval - 10 Sampling interval - 5 Sampling interval - 15 Sampling interval - 5

minutes minutes minutes minutes
Proportional gain - Proportional gain - Proportional gain - Proportional gain -

0.1 0.8 0.5 0.2
Integral time - 0.01 Integral time - 0.01 Integral time - 0.01 Integral time - 0.1
Tank Circular sloping Circular flat bottom Circular flat bottom Circular sloping
Dimension bottom settler settler settler bottom settler
of Final
Clarifiers Centre depth - 3.45m Centre depth - 4.5m Centre depth - 4.5m Centre depth - 5.25m
Sidewall depth - Surface Area - 5,606 Surface Area - 16,120 Sidewall depth -
1.85m m2 m2 1.75m
Surface Area - Surface Area - 6801.0

1,078.0m2 m2
In WWTP A the aerobic suspended activated sludge only treats 83% of the treated primary
effluent. The remaining 17% is treated by the trickling filter with subsequent clarification.
43
Although the sampling campaign did not measure the concentration of the effluent of the
trickling filter process, a “Trickling Filter” unit was included in WWTP A’s model. However,
the performance of the trickling filter will not be calibrated.
The “Trickling Filter” unit has a bed depth of 1.8m and a bed surface of 173m2. The subsequent
clarifier has a flat-bottomed shape, with a surface area of 117m2 and a water depth of 2.5m
(inlet point is 1.25 m from the bottom of the clarifier). The biological model stoichiometry of
the “Trickling Filter” unit was left as its default value in the model. No biological reactions
were enabled in the clarifier.
3.4.1.2.3 Sludge treatment modelling
Although both WWTP B and C have sludge treatment and receives imported sludge from
satellites WWTPs, the flowrate of the imported sludge to WWTP B is absent. Therefore, only
the sludge treatments of WWTP C is modelled in this research.
Sludge storage tank

The sludge imported to, or produced in, WWTP C is stored in the “Plug Flow Tank” unit which
acts as the storage tank, with biological reactions available for modelling the potential
fermentation process (personal communication, S Snowling, Hydromantis, December 2018).
The volume of the tanks is set according to the actual configuration with a universal depth of
5 metres. The volume of the storage tanks is shown in Table 3-4.
44
Table 3-4 The volume of the sludge storage tanks
Tanks Volume, m3
Imported sludge tanks* 17,000
Primary sludge tank 390
SAS tank 250
Raw sludge tank 40,000
Digested sludge tank 4,500
*: The volume of imported sludge tank is calculated based on the historical flow of 857 m3 and the potential
storage time before it reaches WWTP C is estimated as 14 days (personal communication L Wilkinson,
Northumbrian Water, August 2018).
In order to mimic anoxic or anaerobic conditions, the DO of the tanks is set to 0. The “Plug
Flow Tank” unit is originally designed for the suspended growth wastewater treatment and
hence reactive biomasses are assumed present in the tank initially. In this application, in order
to mimic the fermentation that is caused simultaneously by the biomass contained in the sludge
itself rather than by the biomass initially in the tank, the initial concentration of all the biomass
in the “Plug Flow Tank” unit is set to “0” (personal communication, S Snowling, Hydromantis,
December 2018).
Sludge thickening and dewatering

WWTP C has both sludge thickening via drum thickening and sludge dewatering via centrifuge.
GPS-X does include modelling options for such processes, via the “Drum Microscreen” and
the “Hydrocyclone Solid Separation” units, which are corresponding to the actual process used
at WWTP C. In GPS-x, the “Drum Microscreen” and the “Hydrocyclone Solid Separation”
share the same set of operational parameters that control the modelling of thickening and
dewatering, though the default values are different. Considering the parameters will be tested
for sensitivity and adjusted later on, for simplification purpose the “Drum Microscreen” unit is
used for modelling of all thickening and dewatering processes regardless what the actual
process is.
Anaerobic Digestion
In WWTP C, the indigenous primary sludge will be thickened with the imported sludge together.
The indigenous SAS only is thickened alone. The thickened SAS and thickened primary sludge
will be combined and stored. The combined sludge is called Raw sludge. The Raw sludge will
45
then be dewatered and fed to the Thermal Hydrolysis Process (THP). The THP treated sludge
will then be fed to the digester. The digested sludge will be stored in the storage tank prior to
its dewatering.
In this research, the complete model of WWTP C only includes the wastewater treatment, the
thickening of primary sludge and SAS, and the dewatering of raw sludge. The anaerobic
digestion and the subsequent dewatering of the digested sludge are modelled in a separated
layout. The reason for separating the models is that both the TOC:COD ratios and the organic
fraction are found influential to the flow and quality of the produced biogas, and to the soluble
nutrients content of the sludge liquor from the digested sludge dewatering (Hydromantis, 2015).
However, the appropriate TOC:COD ratio fits for the digestion process may not be the same as
the appropriate TOC:COD ratio fits for the modelling of the wastewater treatment. Moreover,
the predicted organic fraction of the dewatered raw sludge generated from the wastewater
treatment model may be substantially to the appropriate organic fraction for the digester feed,
which it is not possible to adjust.
The “Anaerobic Digestion” unit is used to model the sludge anaerobic digestion in WWTP C.
According to the actual configuration, the volume of the digester is set at 18,463 m3 with 1,000
m3 head space. Because there is not enough information, such as the sCOD of sludge treated
before and after the THP process, the THP process is not included in the model.
3.4.2 Influent characterization
The “codstate” model in the “mantis2” library is used for all the wastewater and sludge
modelling. The flow, concentration and fractions of the components of the raw wastewater and
sludge are required to initiate the modelling.
In total, it is necessary to input 66 variables of composite variables, state variables and fractions
in the “codstate” model to thoroughly specify every state variable required by the model
(Hydromantis, 2017). Due to the limited number of parameters actually measured in the
sampling campaign, only flow, COD, TKN and TP, nitrogen-related state variable of NH4+-N
and nitrite and nitrate, phosphorus-related state variable of phosphate, the VSS/TSS ratio, the
organic fractions of COD, the nitrogen fraction, the phosphorus fraction, and the ratios of TOCs
46
to CODs are input to the model. This research assumes that there is no reactive biomass in the
raw wastewater and the variables of store poly-phosphate in PAO inorganic compounds,
organic fractions of volatile fatty acids, methanol and biomasses, inorganic compounds,
inorganic precipitates and soluble gases remain at their default values.
3.4.2.1Influent characterization of wastewater modelling
3.4.2.1.1 Flow
The raw wastewater fed to the wastewater treatment process has been screened and de-gritted,
and hence the screening and de-gritting were not included in the models.
Five sampling campaigns were undertaken for each WWTP. However, some wastewater and
sludge samples were unable to collect in some sampling campaigns. Data from three sampling
campaigns was used for modelling at WWTP A and D, four campaigns for WWTP C, and five
campaigns for WWTP B. The mean flowrate of the raw wastewater of those selected campaigns
will be inputted to the model.
WWTP B and C have sludge return liquor recirculated back to the primary settlement. For
these sites the flow and mass loading from the sludge return liquor is counted in the flow and
concentration of the actual raw wastewater for the purposes of modelling, as per Equation 3-
33 and Equation 3-34 in Section 3.5.6.
3.4.2.1.2 COD and its fractions
The composite variable of total COD is experimentally measured or calculated. The soluble
inert fraction to total COD, readily biodegradable fraction to total COD, and the particulate
fraction to total COD remains as default values of 0.05, 0.20 and 0.13 initially.
The key task of the COD characterization is to match the input sCOD and pCOD to the
experimentally measured values. This is achieved via adjusting the colloidal fraction of slowly
biodegradable COD (frscol), which is a type of sCOD, using Equation 3-11.
47
Equation 3-10
𝑠𝐶𝑂𝐷 − 𝐶𝑂𝐷 × 𝑓𝑟𝑠𝑖 − 𝐶𝑂𝐷 × 𝑓𝑟𝑠𝑠

𝑓𝑟𝑠𝑐𝑜𝑙 =
𝐶𝑂𝐷 × (1 − 𝑓𝑟𝑠𝑖 − 𝑓𝑟𝑠𝑠 − 𝑓𝑟𝑥𝑖)
Where sCOD and COD is experimental measured sCOD and COD, frsi is soluble inert fraction
to total COD, frss is readily biodegradable fraction to total COD, and frxi is the particulate
fraction to total COD.
3.4.2.1.3 Nitrogen compounds and its fractions
TKN and NH4+-N were experimentally measured on all the wastewater samples for the
modelling. Nitrite and nitrate were not experimentally measured on all of the wastewater
samples collected from WWTP B and only for some of the wastewater samples collected from
WWTP C. For those raw wastewater samples for which nitrite and nitrate were not measured
their contents are assumed to be 0 mg/L. Nitrite and nitrate were not measured on the primary
effluent and final effluent from WWTPs B and C. Neither WWTP has SAS recirculated back
to the primary settlement and nitrification does not occur in their aerobic activated sludge
process. Therefore, nitrite and nitrate content of the primary settlement and final clarification
were also assumed to be 0 mg/L.
Soluble TKN was not measured on all raw wastewater samples. The default value of 0.9 is
used.
The N content of the soluble inert material is set as 0.10 for WWTP A, C and D given that there
is a relatively big difference between the observed TKN and NH4+-N in the final effluent
(Roeleveld and van Loosdrecht, 2002, El Sheikh et al., 2016). The default value of 0.05 is set
for WWTP B because difference between the observed TKN and NH4+-N in the final effluent
is small. The N content of inert particulate material is set as 0.03 according to Roeleveld and
van Loosdrecht (2002).
48
3.4.2.1.4 Phosphorus compounds and its fraction
Phosphate was experimentally measured on all the raw wastewater samples. The original
sampling campaign focused on understanding the phosphate flux through the wastewater
treatment works and hence TP was not measured. TP is therefore estimated using COD data.
Both WWTP A and C have historical data for COD and TP of their raw influents. There are
strong positive relationships (p<0.05, examined with Minitab 2018) between TP and COD in
both cases, as shown in Figure 3-1 and 3-2.
WWTP A
12
y = 0.0063x + 1.8753
R² = 0.8198
10
8
TP, mg/L
0
0 200 400 600 800 1000 1200 1400 1600
COD, mg/L
Figure 3-1 Relationship between TP and COD of WWTP C
WWTP C
16
14 y = 0.0163x
12 R² = 0.6979
TP, mg/L
10
8
6
4
2
0
0 200 400 600 800 1000
COD, mg/L
Figure 3-2 Relationship between TP and COD of WWTP C

49
Considering WWTP A and WWTP D both receives domestic wastewater with limited industrial
wastewater and WWTP D is geographically close (approximately 12 miles) to WWTP A
(personal communication, J McCowan, Scottish Water, July 2016), the TP of the raw
wastewater of WWTPs A and D is estimated via COD through the linear regression equation
of TP (mg/L) = 0.0063×COD+1.875 determined by historical data of WWTP A. The same
approach applies to WWTP B and C as they receive substantial amount of industrial wastewater
and are close (approximately 7.5 miles) to each other (personal communication, L Wilkinson,
Northumbrian Water, March 2017). The equation is TP (mg/L) =0.00163×COD.
Both the P content of soluble inert material and inert particulate material are set as 0.005
according to Roeleveld and van Loosdrecht (2002).
3.4.2.1.5 The VSS/TSS ratio
Neither VSS nor TSS were measured on all the wastewater samples. The VSS and TSS values
are not directly input to the model but are calculated by multiplying the inert particulate and
particulate substrate concentrations by a default ratio value of 1.8 g particulate COD/g VSS.
This ratio was not adjusted at any point during the modelling. Because the primary settlement
settles the suspended solids (Tchobanoglous et al., 2003), the VSS/TSS ratio of the inflowing
wastewaters are assumed to be the same as the VS/TS% of their primary sludge.
3.4.2.1.6 The Carbon content in substrates
The Carbon content of soluble inert material, inert particulate material, slowly biodegradable
substrate, soluble substrate, and colloidal material can be defined individually. Similar to the
default setting in GPS-X, for this research a common value was used for all five parameters.
The value given is site specific and is determined by dividing the TOC by the COD of the raw
wastewater.
50
3.4.2.2 Influent characterization of sludge treatment
Only the sludge treatment of WWTP C was modelled in this research. The sludge treatment
prior to anaerobic digestion was modelled in the wastewater treatment model of WWTP C.
Thus, the flow and quality of majority of the sludge streams were predicted generated in the
model simultaneously. However, there were two occasions that the sludge is like the raw
wastewater need to be characterized for their fractions. The first one is the imported sludge that
is combined with the indigenous primary sludge and fed to the primary sludge thickening. The
other is the digester feed sludge.
3.4.2.2.1 Imported sludge
The characterization of the imported sludge focuses on three aspects: COD, TN and TP. The
concentration of each parameter (𝐶𝐼𝑚𝑝𝑆 ) is calculated according to the predicted flow and
quality of the primary sludge and the measured flow and quality of the feed sludge to the
primary sludge drum thickener, and is calculated via Equation 3-11
Equation 3-11
𝑄𝑃𝐹 × 𝐶𝑃𝐹 − 𝑄𝑃𝑆 × 𝐶𝑃𝑆

𝐶𝐼𝑚𝑝𝑆 =
𝑄𝐼𝑚𝑝𝑆
Where 𝑸𝑷𝑭 is the measured flowrate of the feed sludge to the primary sludge thickening, 𝑪𝑷𝑭
is the measured or estimated concentration of COD, TN and TP of the feed sludge to the
primary sludge thickening, 𝑸𝑷𝑺 is the predicted flowrate of the primary sludge, 𝑪𝑷𝑺 is
predicted concentration of COD, TN and TP of the primary sludge, 𝑸𝑰𝒎𝒑𝑺 is measured
flowrate of the imported sludge.
In GPS-X, the TS and VS of the sludge is not possible to predicted by modeling (Hydromantis,
2017). Rather, they are estimated from the COD. By default, the software considers that a gram
of pCOD is equal to 0.56 gram of VSS. Thus, a value of the COD of the digester feed sludge
and the measured VSS/TSS ratio will be input to the “Influent Advisor” in GPS-X and will
then be manually adjusted to a proper value that leads to an adequate TS and VS which are
close to the calculated result.
51
The characterization of the imported sludge is based on the WWTP A and D’s primary sludges
that are co-settled with the indigenous biological sludge recirculated to the primary tanks. This
is because the primary sludge produced in WWTP A and D is actually exported to other
treatment works for thickening and dewatering. The phosphate, ammonium, organic fractions
(such as the inert particulate COD to total COD, the inert soluble COD to total COD, etc.),
nitrogen and phosphorus fraction, and part of the inorganic compounds are the component of
interest to characterize. The values of the components and fractions used for the imported
sludge are taken the mean value of the primary sludge generated from WWTP A and D. Details
are listed in Table 3-5.
Table 3-5 The concentration and fraction used for characterizing the imported sludge
Value of Value of Value
primary primary used for
Parameter Unit
sludge of sludge of Imported
WWTP A WWTP D sludge
ammonia nitrogen gN/m3 72.77 39.85 56.31
ortho-phosphate gP/m3 15.68 10.21 12.95
soluble inert fraction of total COD - 0.00 0.00 0.00
readily biodegradable fraction of total COD - 0.00 0.00 0.00
particulate inert fraction of total COD - 0.21 0.23 0.22
colloidal fraction of slowly biodegradable
- 0.00 0.00 0.00
COD
acetate fraction of total COD - 0.02 0.02 0.02
propionate fraction of total COD - 0.01 0.00 0.01
unbiodegradable cell products fraction of
- 0.02 0.07 0.04
total COD
iron hydroxide gFe(OH)3/m3 311.10 0.00 155.55
iron phosphate gFePO4/m3 444.70 0.00 222.35
magnesium hydrogen phosphate
gMgHPO4/m3 0.49 0.32 0.40
(newberyite)
magnesium ammonium phosphate (struvite) gMgNH4PO4/m3 39.07 0.41 19.74
3.4.2.2.2 Digester feed sludge
The characterization of the imported sludge focuses on four aspect, the COD, TN and TP, and
the flowrate.
In theory, the digester feed sludge comes from dewatered raw sludge, and therefore the COD,
TN and TP mass loading of the two should be the same. Because dewatered sludge is
occasionally imported, and combined with the dewatered raw sludge, the mass loading of the
52
two may be different. However, this is not known for certain because flowrate and quality of
sludge going through the THP process was not measured. If the mass loading of the predicted
dewatered Raw sludge is smaller than the mass loading of the digester feed, then the measured
concentration of the digested feed will be used to characterize the sludge. If the opposite is the
case, the concentration of the digester feed will be estimated from the mass loading of the
dewatered raw sludge via Equation 3-12.
Equation 3-12
𝑄𝐷𝑅𝑆(𝑃)
𝐶𝐷𝐹 = 𝐶𝐷𝑅𝑆 ×
𝑄𝐷𝐹(𝑂)
where 𝑪𝑫𝑹𝑺 is the concentration of COD, TN and TP in the dewatered Raw sludge, 𝑸𝑫𝑹𝑺(𝑷) is
the predicted flowrate of the dewatered Raw sludge and 𝑸𝑫𝑭(𝑶) is the experimentally measured
volume observed of the digester feed.
The observed COD (estimated via VS via Equation 4-4), TN and the TP obtained from the five
sampling campaigns done in this study are 113,137.1 mg/L, 4,315.4 mg/L and 1109.7 mg/L,
respectively. The flowrate is collected from Northumbrian Water and is 876 m3/d (personal
communication, L Wilkinson, Northumbrian Water, 01 March 2017).
The NH4+-N and PO43--P content of the digester feed sludge will be estimated based on the
predicted value of the dewatered raw sludge through the Equation 3-12.
The soluble inert fraction to total COD, readily biodegradable fraction to total COD, the
particulate fraction to total COD, and the colloidal fraction of slowly biodegradable COD are
given values of 0.0, 0.0, 0.20 and 0.0, respectively. Default values were used for all other
fractions and concentrations.
3.4.3 Model calibration and validation
The aim of the calibration process is to match the simulated flow and quality of the effluent
wastewater and sludge to the experimentally measured values, and then to predict the flow and
quality of the unmeasured streams, and finally to study the mass and chemical energy balance
of the WWTPs (Petersen et al., 2002, Hulsbeek et al., 2002).
53
There are two stages in the calibration process, the first is to screen the most influential
parameters that affects the predicted result, and the second stage is to adjust those screened
parameters to a reasonable value that makes the predicted flow and quality of the wastewater
and sludge match experimental observed values.
3.4.3.1 Parameter screening

According to Brun et al. (2002) and Liwarska-Bizukojc et al. (2011), a sensitivity coefficient
(Si,j) is used to demonstrate how influential a parameter is to the predicted result. The Si,j is
calculated via Equation 3-13 .
Equation 3-13
△ 𝑦𝑖 /𝑦𝑖
𝑆𝑖,𝑗 = | |
△ 𝑥𝑖 /𝑥𝑖
where yi is predicted result (output variable), and xi is the input parameter variable which can
be a kinetic of the mechanical settling, mechanical thickening and dewatering, chemical
reaction or biological reaction, a model stoichiometry or a fraction.
Like Liwarska-Bizukojc et al. (2011), a 10% increase is first given to the input parameter
variable xi for obtaining a new predicted result yi. Meanwhile, the △ 𝑥𝑖 /𝑥𝑖 is constant as 0.1.
The Si,j is then calculated. The Si,j is studied on every output variable of the model. If the Si,j is
greater than 0.25, then the input parameter is considered as influential to the predicted
parameter (Liwarska-Bizukojc et al., 2011).
The Si,j is then used to calculate the mean square sensitivity measure δjmsqr. This is used to
demonstrate the influence of a change in a single parameter to the overall predicted results
(Liwarska-Bizukojc et al., 2011). The δjmsqr is calculated using Equation 3-14.
Equation 3-14
𝑛
1
δ𝑗𝑚𝑠𝑞𝑟 2
= √ ∑ 𝑆𝑖,𝑗
𝑛
𝑛=1
Where n is the number of the target variables that characterizes the quality of the wastewater
and sludge.
54
The parameters controlling the settling, biological reactions in the settlers, activated sludge,
sludge storage tanks, anaerobic digestion, and sludge thickening and dewatering will be studied
for both Si,j and δjmsqr. According to Liwarska-Bizukojc et al. (2011), if a parameter has a δjmsqr
greater than 0.1000 or ranks top 10-20th among all the studied parameters, then it will be
selected for adjustment.
3.4.3.2 Parameter adjustment
3.4.3.2.1 Calibration algorithm
GPS-X has an Optimization function to numerically adjust the influential control variable to a
value (within a given range) that makes the predict value of the target variables fit closely to
their observed value (Hydromantis, 2017, Drewnowski et al., 2018). The Nelder-Mead
algorithm is used in Optimization function (Hydromantis, 2017, Drewnowski et al., 2018). The
control variables are commonly kinetics and model stoichiometry of a mechanical process or a
biological reaction.
There are three objective functions in the Optimization function that are used in the calibration
conducted in this research:
1. The “Relative Difference” (Equation 3-15),
2. The “Relative Sum of Squares” (Equation 3-16) and
3. The “Absolute Difference” (Equation 3-17).
Equation 3-15
𝑚 𝑛𝑗
𝐹 = ∑ ∑|𝑧𝑖,𝑗 − 𝑓𝑖,𝑗 |
𝑗=1 𝑖=1
Equation 3-16
𝑚 𝑛𝑗
𝑧𝑖,𝑗 − 𝑓𝑖,𝑗
𝐹 = ∑∑| |
𝑧𝑖,𝑗
𝑗=1 𝑖=1
55
Equation 3-17
𝑚 𝑛𝑗
2
𝑧𝑖,𝑗 − 𝑓𝑖,𝑗
𝐹 = ∑∑( )
𝑧𝑖,𝑗
𝑗=1 𝑖=1
Where zi,j is the measured value of response j in experiment i, fi,j is the value of response
variable j predicted by the process model in experiment i, m is the number of measured
response variables, and the number of experiments for response j.
The “Relative Difference” and “Relative Sum of Squares” aim to minimize the percentage
difference between the predicted values and the observed values of all the target variables.
However, the drawback of these two objective functions is that they prioritize the target
variable that is present at low concentrations as a small numerical change will leads to a big
percentage change. For example, the targeted concentration of A substance is 10 mg/L and or
B substance is 10,000 mg/L. The initial prediction of A and B is 9 mg/L and 10,000 mg/L. The
total relative difference between the initial prediction and the observation is |(9-
10)/10)+(10,000-10,000)/10,000)|=10%. A given change of a kinetic of a mechanical or
biological reaction is able to increase the prediction of A by 1 mg/L and the prediction of B by
10 mg/L. If “Relative Difference” is the objective function, the Optimization will not make the
change of the kinetic because the total numerical difference would be increased to |(10-
10)/10)+(10,000-10,010)/10,000|=0.1%. This also applies to the optimization using “Relative
Sum of Square” as objective function.
The “Absolute Difference” objective aims to minimize the numerical difference between the
predicted values and the observed values of all the target variables. The drawback of this is that
it will prioritize those target variables that have high concentrations (in thousands or tens of
thousands mg/L), since the change in them can be hundreds or thousands mg/L and therefore
leads to a substantial numerical difference. Using the same example stated in last paragraph for
demonstration, the total numerical difference between the initial prediction and the observation
is |(10-9)+(10,000-10,000)|=1 mg/L. A given change of a kinetic of a mechanical or biological
reaction is still able to increase the prediction of A by 1 mg/L and the prediction of B by 10
mg/L. If Absolute Difference is the objective function, the Optimization will not make the
change of the kinetic because the total numerical difference would be increased to |(10-
10)+(10,000-10,010)|=10 mg/L.
56
Regarding the given value range of the control variable, for the biological models the upper
boundary is set as 120% of the default value and the lower boundary is set as 80% of the default
value. If a value range is reported, then the reported range will be used. In some cases, instead
of a value range a single value is reported. If the reported value is higher than 120% or lower
than 80% of the default value, then it becomes the boundary. If the control variable is the
operational efficiency of a mechanical process, the range given is 0-1. The value ranges of the
kinetic and stoichiometry of the biological models are given in Appendix B.
3.4.3.2.2 Calibration procedures
The calibration procedures for the wastewater treatment modelling

The calibration procedures is developed from Fall et al. (2011) and is stepwise. The calibration
procedures of the wastewater treatment are slightly different for WWTP A and D, which have
SAS recirculation to their primary settlers, compared to WWTPs B and C which do not. The
calibration procedures are listed below.
Step 1
If there is substantial change of sCOD and phosphate after primary treatment, then a virtual
unit for sCOD removal and phosphate removal is added between the influent unit and the
primary settlers.
For the virtual sCOD removal, the soluble inert fraction of total COD, colloidal fraction of
slowly biodegradable COD, readily biodegradable fraction of total COD and the chemical
dosage rate of Al2(SO4)3·18H2O are the control variables. The sCOD and pCOD of the raw
wastewater, sCOD and the soluble inert sCOD of the primary effluent, are the target variable.
The soluble inert sCOD, which is not experimentally measured, is set as 90% of the observed
sCOD of the effluent from final clarification (Hulsbeek et al., 2002).
For the virtual ferric dosing for phosphate removal, the chemical dosage rate of
Fe2(SO4)3·9H2O is the control variable and the measured phosphate concentration of the
primary effluent is the target variable.
57
The objective function of the “Optimization” is “Absolute Difference”.
Step 2
For WWTP B and C (o SAS recirculated back to the primary settlement) the influential
parameters on the primary settlers are chosen first. The target variables of this tuning are the
pCOD, sCOD, TKN of the primary effluent and the TSS, VSS, TN and TP of the primary
sludge. The objective function of the Optimization is “Relative Sum of Squares”.
The influential parameters of the activated sludge and final settlers are then chosen for the
adjustment. However, if the influential parameter is from the ammonium oxidizers group or
the nitrite oxidizer group of the biological model, or is only influential (with a S i,j > 0.25) to
the nitrogen derivatives of ammonium, nitrite and nitrate, then it is not adjusted. This was the
approach taken by (Fall et al., 2011). The target variables are the pCOD, sCOD, TKN, and
phosphate of the final effluent and the flow, TSS, VSS, TN and TP of the WAS. The objective
function of the “Optimization” is “Relative Sum of Squares”.
For WWTP A and D (SAS recirculated back to the primary settlement) the influential
parameters of the primary settlers, activated sludge, and final settlers are chosen for calibration
all at onceat this step.The target variables are the pCOD, sCOD, TKN, and phosphate of the
primary effluent, the pCOD, sCOD and phosphate of the final effluent, the TSS, VSS, TN and
TP of the primary sludge and SAS, and the flow of the SAS. The objective function of the
“Optimization” is “Relative Sum of Squares”.The same approach is taken for the ammonium
and nitrite oxidizer groups as for WWTPs B and C.
Step 3
The nitrogen related parameter will be adjusted. The target variables are the pCOD, sCOD,
TKN, and phosphate of the primary effluent, the pCOD, sCOD, ammonium, TKN, Nitrate, TN,
and phosphate of the final effluent, the TSS, VSS, TN and TP of the primary sludge and SAS,
and the flow of the SAS. The objective function of the “Optimization” is “Relative Sum of
Squares”.
58
The calibration procedures for the sludge thickening and dewatering treatment prior to
anaerobic digestion
The influential parameters of the biological model of the sludge storage tank and sludge
thickening and dewatering will be adjusted at the same time. The target variables are commonly
the COD, sCOD, pCOD, ammonium, TKN and phosphate of the filtrate and centrate and the
COD, TN and TP of the thickened and dewatered sludge. In the case of the raw sludge
dewatering, the COD, TN and TP of the stored sludge (prior to dewatering) will also be chosen
as target variables in the tuning. That is because the stored sludge was what was sampled and
measured in the sampling campaign. The objective function of the “Optimization” is “Relative
Difference”.
The calibration procedures for anaerobic treatment and subsequent digestion

The influential parameters for the biological model as it relates to the anaerobic digestion and
dewatering of digested sludge are adjusted at the same time. The fraction of inflowing sludge
and the ratio of TOC content to per gram particulate CODs are also adjusted at this time. The
target variables are the TOC content of the feed sludge, the flow of the produced CO2, CH4 and
total biogas, the COD, sCOD, pCOD, ammonium, TKN and phosphate of the filtrate and
centrate, and the COD, TN and TP of both the stored and dewatered digested sludge. The
objective function of the “Optimization” is “Relative Difference”.
3.4.3.3 Validation
Rieger et al. (2012) considers that it is acceptable to have ± 5-15 % difference between the
predicted data and the observed data. However, current literatures rarely reported a quantitative
method for validating predicted data generated from the steady state modeling to the observed
data. But Liwarska-Bizukojc et al. (2011) reported a quantitative method for validating the
goodness of fit between the observed values and the predicted values produced in the dynamic
modelling. The method is to calculate the average relative difference by comparing a set of
predicted values to the corresponding observed values via Equation 3-18.
59
Equation 3-18
𝑂𝑉 − 𝑃𝑉
∑𝑛𝑛=1(|
𝐴𝑅𝐷 = 𝑂𝑉 | × 100%)
𝑛
where OV is the observed value, PV is the predicted value, n is the number of target variables
selected in the optimization.
In the example of Fall et al. (2011) which reported both the observed value and the predicted
values generated from steady state modeling, the ARD calculated is 23%. For this research an
ARD of no greater than 15% was deemed acceptable, to be consistent with the range of values
observed by Rieger et al. (2012).
Furthermore, a visualization aid developed by Fall et al (2011) was also used to demonstrate
the goodness of fit of the predicted values. The method is to use the observed value (in y-axis)
to plot against the predicted value. A diagonal line with a slope of 1 is drawn on all plots to use
as a reference line. The graphs are produced via Matlab 2017.
3.5 Construction of mass and chemical energy balances
The wastewater and sludge treatment in the investigated WWTPs consists of primary
settlement, aerobic activated sludge as secondary biological treatment, secondary clarification,
sludge storage and subsequent thickening and dewatering, and anaerobic digestion. This study
firstly examines the mass and energy balance of each individual process and then investigates
the balances of the entire WWTPs.
3.5.1 Calculation of mass loading
Mass loading within wastewater is calculated by multiplying flow (Q, m3/day) by concentration
(C, mg/L) in a sample collected simultaneously to the flow measurement (Equation 3-19).
Mass load is reported in units of kg/day or tonne/day, as appropriate.
Equation 3-19
𝑀=𝑄 × 𝐶
Where Q is the flow of the wastewater, and C is the concentration of a given contaminant.
60
The concentration (C%) of sludge is expressed as percentage weight in dried solid and is
therefore a weight to weight unit. Thus, the mass loading within sludge is calculated by
multiplying the flow by the specific density (ρ, ton/m3) of sludge first, and then by its TS% and
the corresponding concentration, as shown in Equation 3-20.
Equation 3-20
𝑀 = 𝑄 × 𝜌𝑆𝑙𝑢𝑑𝑔𝑒 × 𝑇𝑆% × 𝐶%
Where Q is the flow of the sludge, 𝝆𝑺𝒍𝒖𝒅𝒈𝒆 is the specific density of sludge, TS% is the dried
solid percentage of the sludge, and C% is the percentage concentration to the dried solid.
However, the observed sludge concentration is to serve the modelling. In the modelling
software GPS-X, the component concentration is expressed weight to volume (w/v)
concentration in unit of mg/L. Therefore, the w/w concentration has to be converted in to mg/L
concentration.
In the sludge density calculation shown in Appendix C, the sludges sampled have a TS% varies
from 0.5% to 24.0%, and VS/TS% varies from 60% to 80%. In this range of TS% and VS/TS%,
the specific density of sludge falls in the range of 1,001 to 1,061 g/L. Thus, the specific density
of all the sludges in this study were assumed as 1,000 g/L. A 1,000g/L sludge specific density
is also applied in Lu (2006).
Therefore, the w/v TS content of the sludge (TSs, g/L) is calculated via Equation 3-21.
Equation 3-21
𝑇𝑆𝑠 = 𝑇𝑆% ∗ 1,000
For other sludge component of VS, TOC, TN and TP, their w/v concentration is calculated via
multiplying the TSs with their weight percentage to TS. For example, the w/v VS content of
the sludge is calculated by multiplying the TS with VS/TS%.
3.5.2 Calculation of chemical energy loading
As chemical energy content of wastewater (Eww, kJ/L) is also a concentration, chemical energy
(E) loading within wastewater is calculated by multiplying flow with the corresponding Eww
(as shown in Equation 3-22).
61
Equation 3-22
𝐸 = 𝑄 × 𝐸𝑤𝑤
where Q is the flow of the wastewater and Eww
Since chemical energy content (Eds, kJ/g) of sludge is the calorific value of the dried solid,
chemical energy loading within the sludge is calculated by multiplying flow by the specific
density (ρ, ton/m3) of sludge first, and then by its TS% and the corresponding Eds (as shown in
Equation 3-23).
Equation 3-23
𝐸 = 𝑄 × 𝜌𝑆𝑙𝑢𝑑𝑔𝑒 × 𝑇𝑆% × 𝐸𝑑𝑠
where Q is the flow of the wastewater, ρSludge is the density of sludge and is assumed 1,000g/L,
TS% is the dried solid percentage of the sludge, and Eds is the chemical energy content (kJ/g)
of the dried sludge.
The chemical energy loading may also be estimated through the relationship between the
chemical energy content (in kJ/L or kJ/g) and the TS, VS and COD determined in Chapter 4.
3.5.3 Flow balance
As shown in Section 3.5.1 and 3.5.2 flow is the key element for acquiring mass and energy
loading. In this study, flow of all the raw inlet and SAS sludge, feed sludge to thickening and
dewatering process, and in-flowing and out-flowing sludge to the digester were measured and
collected by water companies (stated in Section 3.1.3). However, the flow of primary effluent
of WWTP A, B and D, secondary effluent of WWTP A and D, the primary sludge of WWTP
B and C, and the out-flowing sludge of WWTP B and C’s sludge thickening and dewatering
were not measured by water companies. Neither was it possible to measure these variables with
a portable flow meter. Given that the total inlet flowrate should equal the total outlet flow (Puig
et al., 2008), in this study the effluent wastewater flow rate is therefore calculated using
Equation 3-24.
Equation 3-24
𝑄𝐼𝑛 = 𝑄𝐸𝑓𝑓 + 𝑄𝑆
where Qin is the flow of all influents, QEff is the flow of all the effluent wastewater streams, and
QS is the flow of all the effluent sludge streams. The influent liquid streams, effluent wastewater
62
streams and sludge streams of the flow balance of each individual process are listed in Table
3-6.
Table 3-6 The influent liquid streams, effluent wastewater streams and sludge streams of the flow balance of
each individual process
Process QIn QEff QS

WWTP A and D:
Raw wastewater plus
recirculated biological sludge
Primary Primary effluent Primary sludge
WWTP B and C:
Raw wastewater plus sludge
return liquor
Primary effluent
Activated Sludge Mixed liquor N/A
RAS
RAS
Secondary Clarifier Mixed liquor Final Effluent
SAS
Activated Sludge +
Primary effluent Final Effluent SAS
Secondary Clarifier
WWTP A and D: WWTP A and D:
Raw wastewater Primary Sludge
Whole Wastewater
Final Effluent
treatment work WWTP B and C: WWTP B and C:
Raw wastewater and sludge Primary Sludge +
return liquor SAS
Sludge Holding Tank Sludge N/A Stored sludge

Sludge Thickening and Thickened or
Sludge Return Liquor
Dewatering Dewatered sludge
Anaerobic Digestion Sludge feed N/A Digested Sludge
Wastewater treatment + Raw wastewater and sludge Final Effluent

Digested Sludge
Sludge Treatment return liquor Return liquor
3.5.4 Mass balance
This study investigates the mass balances of COD, TN and TP of all the processes and the mass
balances of TS and VS of the sludge thickening and dewatering processes. In wastewater and
sludge treatment there are commonly multiple outflows: (1) treated effluent, (2) sludge and (3)
63
gaseous emissions (captured or otherwise). Therefore, the mass balance of a process is
expressed as shown in Equation 3-25.
Equation 3-25
𝑀𝐼𝑛 = 𝑀𝐸𝑓𝑓 + 𝑀𝐺 + 𝑀𝑆
Where MIn is the influent mass, MEff is the effluent mass in the treated effluent, MG is the
effluent mass in the gaseous emission, and MS is the effluent mass in the removed sludge. The
influent liquid streams, effluent wastewater streams and sludge streams used in the mass
balance of each individual process are listed in Table 3-7.
Table 3-7 The influent and effluent wastewater streams and sludge streams used in the mass balance of each
individual process
Process MIn MEff MG MS

WWTP A and D:
Raw wastewater
and recirculated biological
sludge
Primary Settlement Primary effluent N/A Primary Sludge
WWTP B and C:
Raw wastewater
and sludge return liquor
Primary effluent
Activated Sludge Mixed liquor CO2 and N2 N/A
RAS
Secondary RAS
Mixed liquor Final Effluent CO2 and N2
Clarification WAS
Activated Sludge +
Primary effluent Final Effluent CO2 and N2 WAS
Secondary Clarifier
Whole Wastewater
Final Effluent CO2 and N2
Raw wastewater Primary Sludge +
and sludge return liquor WAS
Sludge Holding
Sludge N/A CO2 and N2 Stored sludge
Tank
Thickened and
Sludge Thickening
Sludge Return Liquor N/A Dewatered
and Dewatering
sludge
Biogas
Anaerobic Digestion Sludge feed N/A (CH4 and Digester Sludge
CO2)
Wastewater
Raw wastewater Final Effluent Biogas, CO2
treatment + Sludge Digester Sludge
and sludge return liquor Return liquor and N2
Treatment
64
To investigate the distribution pattern of COD, TN and TP in individual process, in Chapters 4
and the proportionate mass flows of substances as treated effluent, sludge, or gaseous emission
are expressed as percentages of the total influent mass.
If both flow and concentration of all inflow and outflow streams are directly measured or
modelled, the mass balance is considered as closed. Nevertheless, as there may be error in both
flow and concentration measurement it is not always the case that inflowing mass is exactly
equalled by mass flows out. The error of the closed mass balance is studied by comparing the
numerical difference between the right and the left of Equation 3-26. Considering flow
measurement may be subject to errors of ± 15% (Agu et al., 2017), if the difference is within ±
15%, the mass balance is considered as balanced and correct. If the difference is outside of ±
15%, the mass balance is considered as incorrect, and discussion is provided on the possible
causes of the inaccuracy.
If an individual mass loading is calculated by directly subtracting the sum of all out-flowing
mass loadings from the inflows (as shown in Equation 3-26), that mass balance is always
balanced. This type of mass balance is considered as open mass balance. Its accuracy is
examined via comparing the percentage mass loading of each component to the theoretical or
reported value in literature.
Equation 3-26
𝑀𝑂𝑢𝑡(𝑢𝑛𝑘𝑛𝑜𝑤𝑛) = 𝑀𝐼𝑛 − ∑ 𝑀𝑂𝑢𝑡(𝑘𝑛𝑜𝑤𝑛)
Where MOut(unknown) is the effluent mass which is neither measured or calculated, MIn is the
influent mass, and ∑ 𝑴𝑶𝒖𝒕(𝒌𝒏𝒐𝒘𝒏) is the sum of measured or calculated effluent mass. This
method has mainly been applied to the activated sludge and its subsequent final clarification
and sludge storage. This is because flows and concentrations of their potential gaseous
products were not measured.
3.5.5 Chemical energy balance
If there is no chemical energy consumed during the process then the chemical energy balance
is expressed as the Equation 3-27.
65
Equation 3-27
𝐸𝐼𝑛 = 𝐸𝐸𝑓𝑓 + 𝐸𝐺 + 𝐸𝑆
Where EIn is the influent chemical energy, EEff is the effluent chemical energy in the treated
effluent wastewater, EG is the effluent chemical energy in the gaseous emission, and ES is the
effluent chemical energy in the sludge retained in the treatment process. Although gaseous
products are generated in some processes, if the gaseous product (such as CO2 and N2) has a
calorific value of zero then they are not accounted for in the energy balance calculation. The
influent liquid streams, effluent wastewater streams and sludge streams of the energy balance
of each individual process are listed in Table 3-8.
Table 3-8 The influent and effluent wastewater streams and sludge streams of the energy balance for each
individual process
Process EIn EEff EG ES

WWTP A and D:
Raw wastewater
and recirculated biological
sludge
Primary Settlement Primary effluent N/A Primary Sludge
WWTP B and C:
Raw wastewater
and sludge return liquor
Primary effluent
Activated Sludge Mixed liquor N/A N/A
RAS
Secondary RAS
Mixed liquor Final Effluent N/A
Clarification WAS
Activated Sludge +
Primary effluent Final Effluent N/A WAS
Secondary Clarifier
Whole Wastewater
Final Effluent N/A
Raw wastewater Primary Sludge +
and sludge return liquor WAS
Sludge Holding
Sludge N/A N/A Stored sludge
Tank
Thickened or
Sludge Thickening
Sludge Return Liquor N/A Dewatered
and Dewatering
sludge
Biogas
Anaerobic Digestion Sludge feed N/A Digester Sludge
(CH4)
Wastewater
Raw wastewater Final Effluent Biogas
treatment + Sludge Digester Sludge
and sludge return liquor Return liquor (CH4)
Treatment
66
The biological elements of secondary treatment and anaerobic digestion may consume part of
the influent chemical energy. Specifically, the difference between the in-flowing energy and
the out-flowing energy is considered as loss of energy. Their energy balance is expressed as
Equation 3-28.
Equation 3-28
𝐸𝐼𝑛 = 𝐸𝐸𝑓𝑓 + 𝐸𝐺 + 𝐸𝑆 + 𝐸𝐿𝑜𝑠𝑠
where EIn is the influent chemical energy, EEff is the effluent chemical energy in the treated
effluent wastewater, EG is the effluent chemical energy in the gaseous emission, ES is the
effluent chemical energy in the removed sludge, and ELoss is the difference between the in-
flowing chemical energy and the out-flowing chemical energy.
If the chemical energy loading of all the inflows and outflows are calculated by multiplying
flow and concentration the balance is considered as closed. The examination of its accuracy is
described in Section 3.5.4.
3.5.6 Calculations related to the primary treatment
In the construction of the balances the major complexity is within the primary settlement, as
the influent not only consists of raw inlet wastewater but also of recirculated biological sludge,
or sludge return liquor. Therefore, two issues need resolving: the first is to calculate the actual
primary sludge produced due to the sedimentation of the particulates from the raw wastewater,
and the second is to estimate the flow of the sludge return liquor as it is not measured by the
water companies.
For the first issue, this study assumes that all of the biological sludge recirculated back to the
primary settlement will completely settle and be completely discharged within the primary
sludge. The mass of actual primary sludge is therefore calculated by subtracting the mass of
biological sludge from primary sludge in the mass balance, as indicated in Equation 3-29. An
equivalent expression applies to energy (Equation 3-30) and nutrients.
67
Equation 3-29
𝑀𝑅𝐼 = 𝑀𝑃𝐸 + 𝑀𝐴𝑃𝑆 = 𝑀𝑃𝐸 + (𝑀𝑃𝑆 − 𝑀𝐵𝑆 )
Where MRI is the influent mass in the raw inlet, MPE is the effluent mass within the primary
effluent, MAPS is the is effluent mass of the actual primary sludge, MPS is effluent mass within
the primary sludge combined with actual primary sludge and biological sludges, and MBS is the
mass of all the biological sludge recirculated back to the primary settlement tank.
Equation 3-30
𝐸𝑅𝐼 = 𝐸𝑃𝐸 + 𝐸𝐴𝑃𝑆 = 𝐸𝑃𝐸 + (𝐸𝑃𝑆 − 𝐸𝐵𝑆 )
Where ERI is the influent chemical energy with in raw inlet, EPE is the effluent chemical energy
within the primary effluent, EAPS is the is effluent chemical energy of the actual primary sludge,
EPS is effluent chemical energy within the primary sludge combined with actual primary sludge
and biological sludges, and EBS is the chemical energy of all the biological sludge recirculated
back to the primary settlement.
For the second issue the flow of the return liquor is estimated through the flow balance and the
TS mass balance. The flow and TS mass balance of a sludge thickening or dewatering facility
can be described as in Equations 3-31 and 3-32.
Equation 3-31
𝑄𝑊𝑆 = 𝑄𝑅𝐿 + 𝑄𝐷𝑒𝑤𝑆
Where 𝑸𝑾𝑺 is the influent wet sludge flow, 𝑸𝑹𝑳 is the flow of the return liquor (the removed
wastewater), and 𝑸𝑫𝒆𝒘𝑺 is the flow of the effluent thickened or dewatered sludge.
Equation 3-32
𝑇𝑆 𝑇𝑆 𝑇𝑆
𝑀𝑊𝑆 = 𝑀𝑅𝐿 + 𝑀𝐷𝑒𝑤𝑆
Where 𝑴𝑻𝑺 𝑻𝑺
𝑾𝑺 is the influent TS mass, 𝑴𝑹𝑳 is the TS mass within the return liquor (the removed
wastewater), and 𝑴𝑻𝑺

𝑫𝒆𝒘𝑺 is the TS mass within thickened or dewatered sludge.
68
According to Equation 3-20 and 3-31, the Equation 3-32 is derived into Equation 33 as
Equation 3-33
𝑇𝑆
𝑄𝑊𝑆 × 𝑇𝑆𝑊𝑆 % = 𝑄𝑅𝐿 × 𝐶𝑅𝐿 + (𝑄𝑊𝑆 − 𝑄𝑅𝐿 ) × 𝑇𝑆𝐷𝑒𝑤 %
𝑄𝑊𝑆 × (𝑇𝑆𝑊𝑆 % − 𝑇𝑆𝐷𝑒𝑤𝑆 %)

𝑄𝑅𝐿 = 𝑇𝑆
( 𝐶𝑅𝐿 − 𝑇𝑆𝐷𝑒𝑤𝑆 %)
where 𝑸𝑹𝑳 is the flow of the return liquor (the removed wastewater), 𝑸𝑾𝑺 is the influent wet
sludge flow, 𝑻𝑺𝑫𝒆𝒘𝑺 % is the TS% of the thickened or dewatered sludge, 𝑻𝑺𝑾𝑺 % is the TS%
of the wet sludge, and 𝑪𝑻𝑺
𝑹𝑳 is the TS concentration of the return liquor. Because this study
considers the specific density of the sludge to be 1 g/L of the sludge, 1% TS% is equal to 10,000
mg/L of TS.
Once the flow of the sludge return liquor is known, the concentration of the actual in-flowing
wastewater containing sludge return liquor can be calculated using Equation 3-34.
Equation 3-34
𝑀𝑅𝐼 + ∑ 𝑀𝑆𝑅𝐿
𝐶𝐼𝑛(𝑎𝑐𝑡𝑢𝑎𝑙) =
𝑄𝐼𝑛(𝑎𝑐𝑡𝑢𝑎𝑙)
Where MRI is the mass within the raw inlet, ∑ 𝑀𝑆𝑅𝐿 is the sum of mass with the sludge return
liquor, QIn(actual) is the actual flow which includes all the raw wastewater, filtrates and centrates
to primary settlement.
69
Chapter 4 Results and Discussion: Chemical energy of wastewater and
sludge
This chapter discusses the chemical energy within wastewater and sludge. It examines the
performance of an innovative wastewater drying method developed in this research (in Section
4.1.1). It presents the energy-related characteristics of wastewater (in Section 4.1.2) and sludge
(in Section 4.2.1), and then investigates the relationships between routinely measured variables
and wastewater energy (in Section 4.1.3 and 4.1.4) or sludge energy (in Section 4.2.2).
4.1 Chemical energy of wastewater
4.1.1 Method Improvement
A key objective of this research was to develop a more reliable method for the routine
determination of municipal wastewater energy content. The method should be effective in
retaining the COD in the dried wastewater because any loss of COD will result in loss of energy
and therefore an erroneously low value for energy content will result (Heidrich et al., 2010).
Moreover, the method should be also time efficient since the current measurement could take
approximately weeks to finish (Heidrich et al., 2010).
COD loss is primarily associated with the sample drying process (Heidrich et al., 2011). In this
study, the centrifugal evaporation is used to dry the wastewater and therefore the effect of
drying by centrifugal evaporation was tested. Results indicated that the centrifugal evaporation
drying of samples has a greater COD recovery (mean 84.8%, n=36) than both freeze-drying
(73.8 – 81.7%) and oven-drying (51.3 – 56.2%) (Heidrich et al., 2010). Using the centrifugal
evaporation drying process, there was no particular difference in COD loss during drying of
raw wastewater, primary treated and secondary treated, despite the large differences in initial
COD: 87.4% (n=16), 82.6% (n=12) and 82.9% (n=8), respectively. Losses of COD during
drying have previously been ascribed to loss of volatile organic compounds such as
acetate (Heidrich et al., 2010), and that may be the case with losses during centrifugal
70
evaporation drying also. Nevertheless, centrifugal evaporation still incurs lower losses than the
freeze-drying process.
In terms of the drying time, centrifugal evaporation is able to simultaneously dry six 400 mL
wastewater samples typically less than 72 hours. Comparing to freeze-drying which requires
28 days to dry 1.5 L wastewater sample (Heidrich et al., 2010), centrifugal evaporation is a far
quicker drying process but does not compromise the effectiveness in retaining COD.
4.1.2 Wastewater characteristics
Summary statistics for the COD and energy content of the wastewaters (spot raw, composite
raw, composite primary treated and composite final effluent) of all four study sites are shown
in Table 4-1.
The CODs of raw wastewaters A, B and C are significantly higher than that of wastewater D
(p < 0.05), and the COD of Wastewater B is statistically higher than Wastewater C (p < 0.05).
Notwithstanding these differences, the four domestic wastewaters sampled are typical of COD
concentrations of domestic wastewaters internationally (430 mg/L to 800 mg/L for medium
and high strength wastewaters respectively) (Tchobanoglous et al., 2003). Unsurprisingly,
given that organic matter is removed through the treatment process, there is a consistent
decrease in COD, at all wastewater treatment plants, from raw wastewater to final effluent.
The overall mean energy content of the spot raw and composite raw wastewaters falls in the
previously reported range of 1.12 kJ/L and 41.11 kJ/L (Korth et al., 2017, Heidrich et al., 2010,
Shizas and Bagley, 2004). Similar to the COD, there is a consistent decrease in chemical energy
content of wastewater (Eww, kJ/L), at all wastewater treatment plants, from raw wastewater to
final effluent, with the overall mean for the latter being < 1.5 kJ/L.
71
Table 4-1 Mean COD and Eww (± standard deviation) of wastewater samples from this research compared to
literature values
Composite
Composite Composite
Parameter Spot Raw Primary
Raw Final Effluent
Treated
Centrifugal Vacuum Dried
530.1 ±276.3 758.0 ±433 312.3 ±158.0 38.24 ±10.07
WWTP A COD, mg/L
(n=12) (n= 5) (n=5) (n=5)
10.20 ±4.20 12.22 ±8.30 4.65 ±2.83 0.14 ±0.09
Eww, kJ/L
(n11) (n=5) (n=5) (n=5)
714.7 ±229.1 613.1 ±57.7 392.2 ±40.7 57.5 ±28.4
WWTP B COD, mg/L
(n=12) (n=5) (n=5) (n=5)
11.94 ±4.70 11.00 ±1.57 6.39 ±0.71 0.54 ±0.25
Eww, kJ/L
(n=12) (n=5) (n=5) (n=5)
568.9 ±148.1 522.8 ±109.3 450.0 ±96.2 93.8 ±27.4
WWTP C COD, mg/L
(n=12) (n=9) (n=5) (n=5)
6.85 ±2.77 7.86 ±2.63 6.87 ±2.15 1.31 ±0.16
Eww, kJ/L
(n= 12) (n=9) (n=5) (n=3)
409.6 ±189.8 305.4 ±113.5 150.0 ±55.9 24.89 ±6.75
WWTP D COD, mg/L
(n=12) (n=5) (n=5) (n=5)
6.53 ±6.08 4.40 ±1.01 2.17 ±0.96 0.11 ±0.10
Eww, kJ/L
(n=11) (n=5) (n=5) (n=4)
555.8 ±236.3 545.4 ±251.4 326.1±147.0 51.51 ±31.48
Overall COD, mg/L
(n=48) (n=24) (n=20) (n=19)
8.90 ±4.98 8.70 ±4.88 5.02 ±2.56 0.46 ±0.48
Eww, kJ/L
(n=46) (n=24) (n=20) (n=17)
Oven Dried
Heidrich et al. 647.3 ± 100.6
COD, mg/L
(2010) (n=2)
6.95 ± 1.91
Eww, kJ/L
(n=2)
Korth et al. 1275 ± 915
COD, mg/L
(2017) (n=17)
6.30 ± 4.19
Eww, kJ/L
(n=23)
Shizas and
COD, mg/L 431(n= 1)
Bagley (2004)
Eww, kJ/L 6.30 (n= 1)
Freeze Dried
Heidrich et al. 647.3 ± 100.6
COD, mg/L
(2010) (n= 2)
12.20 ± 6.51
Eww, kJ/L
(n= 2)
Korth et al. 1817 ± 1091
COD, mg/L
(2017) (n= 3)
20.22 ± 10.65
Eww, kJ/L
(n= 6)
72
4.1.3 The relationship between COD and chemical energy content of wastewater
For the purposes of data analysis samples were grouped into four categories: (1) composite raw
wastewater samples only (2) all composite samples i.e. raw, settled and final effluent (3)
individual spot raw wastewater samples collected at different times through a day, and (4) all
wastewater samples (both composite and spot samples).
The tested variables (stated in Section 3.2.1) were first analyzed for their correlation with Eww.
For the composite raw samples COD, pCOD, TOC, DOC and VS all had a strong correlation
with Eww (Spearmen Rho >0.85, p<0.05). For all composite samples COD, pCOD, TOC, POC,
Tce and chemical energy content of dried wastewater (Edww, kJ/g) all have a strong correlation
with Eww (Spearmen Rho >0.90, p<0.05). For the spot raw samples COD and Edww have the
strongest correlation with Eww (Spearmen Rho >0.70, p<0.05), and for all composite and spot
samples combined COD, pCOD and Edww all have strong positive correlations with Eww
(Spearmen Rho >0.85, p<0.05). COD has a ubiquitously strong correlation with Eww,
irrespective of sample group.
The variables with the strongest correlation with the Eww were analysed by the Best Subset
function of Minitab to identify the most appropriate variables (in terms of r2 and the adjusted
r2 (r2adj)) to use in a regression model. Although an increase in the number of variables might
improve the value of r2, if r2adj is not simultaneously increased, then the variable tested could
be excluded in the regression (Minitab Inc., 2017b). The best regression model was considered
that having the highest r2adj but with as few variables as possible. The results of the Best Subset
study are shown in Table 4-2.
73
Table 4-2, Best regression for estimating wastewater energy
No. of Variables r2 r2adj Parameter Required

Composite raw, n=24
1 88.9 88.4 pCOD
1 86.5 85.9 COD
1 83.7 83.0 TOC
2 91.1 90.2 pCOD, DOC
3 91.9 90.7* COD, pCOD, DOC
4 92.1 90.4 COD, pCOD, DOC, VS
5 92.1 89.9 COD, pCOD, TOC, DOC, VS
All composite, n=59
1 92.9 92.8 COD
1 92.7 92.5 pCOD
1 91.4 91.2 TOC
2 93.9 93.7 COD, pCOD
3 95.0 94.8* pCOD, TOC, VS
4 95.2 94.8 COD, pCOD, TOC, VS
5 95.3 94.8 COD, pCOD, TOC, VS, TScve
6 95.3 94.8 COD, pCOD, TOC, VS, TScve, Edww
Spot raw, n=46
1 51.3 50.1 Edr
1 50.8 49.7 COD
1 41.1 39.8 TScve
2 89.7 89.2 TScve, Edww
3 90.8 90.1* sCOD, TScve, Edww
4 90.8 90.0 COD, sCOD, TScve, Edr
All samples, n=107
1 77.2 77.0 COD
1 68.3 68.0 Edww
1 52.7 52.2 TScve
2 90.9 90.7 TScve, Edww
3 91.5 91.2* COD, TScve, Edww
4 91.5 91.2 COD, P-PO43-, TScve, Edww
5 91.5 91.1 COD, sCOD, P-PO43-, TScve, Edww
2
*: The regression with highest r adj but as few variables as possible
For both the composite raw and composite of all samples, the r2 of all the single or multiple
regression analyses varies within a narrow range 88.9%-92.1% and 92.9%-95.3%, respectively.
For the composite raw, the r2 of the regression determined solely by COD, pCOD or TOC with
Eww, of the composite raw samples are all above 80.0 % (p<0.05). For all composite samples
74
the r2 values for the regression of COD, pCOD or TOC with E ww are over 90% in all cases
(p<0.05).
In both cases of the spot raw and all samples, the best regression comprises the TScve and Edww.
In fact, Eww is calculated by multiplying TScve and Edww together. If both TScve and Edww are
known, there is no need to use regression to estimate Eww. For both sample groups, regressions
determined solely by COD still result in the strongest single variable regressions.
Regression based solely on COD always ranks in the top 2 of single variable regressions for
each sample group, and the r2 of regression determined by pCOD or TOC of the composite
samples is always greater than 90%. Hence, COD is selected to plot against the Eww of different
sample groups, and pCOD and TOC are chosen to plot against the Eww of the composite samples.
The plot, regression line (red solid line), the confidence interval (green dash line) and the
prediction interval (purple dash line) of each individual regression are shown in Figure 4-1.
The regression equations of each individual regression are listed in Table 4-3. It is worth noting
that the Minitab 18 can only display the confidence interval and prediction interval on
regressions has intercept in its regression equation. However, in 5 out of 6 cases shown in
Figure 4-1, the intercept is not statistically significant (p>0.05) and can be potentially removed
from the regression equation (Minitab Blog Editor, 2013). Thus, the regression equation with
no intercept is also worked out and listed in the Table 4-3. Although the regression with no
intercept has a higher r2 than the regression with intercept does, That is because of the r2 is
calculated differently and it does not mean the quality of the former is better than the latter
(Minitab Inc., 2017a).
75
(A) (B)
(C) (D)
(E) (F)
Figure 4-1 Plots of Eww of different sample groups against COD, pCOD or TOC. (A) E ww of Composite Samples
against pCOD, (B) Eww of Composite Samples against TOC, (C) Eww of Composite Raw against COD, (D) Eww
of Composite Samples against COD; (E) Eww of Spot Raw against COD, and (F) Eww of All Samples against
COD
Table 4-3 Regression equations of regressions between Eww and pCOD, TOC and COD shown in Figure 4-1
Regression Equation without
n Regression Equation with intercept*
intercept*
Figure 4-1 (A) 60 Eww=0.172+0.0227 pCOD(r2=92.7%) Eww=0.0231 pCOD (r2=96.7%)
Figure 4-1 (B) 61 Eww= -0.560+0.0599 TOC(r2=91.7%) Eww=0.0562 TOC (r2=96.0%)
Figure 4-1 (C) 24 Eww= -1.149+0.0181 COD(r2=86.5%) Eww=0.0163 COD (r2=96.6%)
Figure 4-1 (D) 61 Eww= -0.549+0.0171 COD(r2=93.2%)** Eww=0.0161 COD (r2=96.7%)
Figure 4-1 (E) 46 Eww= 0.02+0.0155 COD(r2=50.9%) Eww=0.0156 COD (r2=88.5%)
Figure 4-1 (F) 107 Eww= -0.382+0.0164 COD(r2=77.2%) Eww=0.0158 COD (r2=91.6%)
*: Eww is unit of kJ/L. pCOD, TOC and COD are in unit of mg/L.
**: The intercept of the regression equation is found statistically significant.
76
In Figure 4-1 the sample sizes of the 6 datasets shown vary from 24 to 107, compared to a
sample size of only 19 in the work of Korth et al (2017). Moreover, apart from the regression
determined by COD for the spot raw sample (Figure 4-1 (C)), the r2 of the regression models
shown in Table 4-3 are also stronger than those reported for the COD (n=14, r2=65%) and TOC
(n=19, r2=66%) models in Korth et al. (2017). Since the univariate regressions determined by
the single variables of COD, pCOD or TOC are almost as strong as the multivariate regression,
using these individual variables appears to be a very reliable means of estimating the inflowing
chemical energy to a WWTP.
The regression equations shown in bottom 4 rows of Table 4-3 indicate the numerical
relationship between the Eww and COD. Because the Eww is in unit of kJ/L and the COD is in
unit of mg COD/L, the slope will be in unit of kJ/mg COD. If the slope is simply considered
as how many kJ of chemical energy is contained in 1 mg of COD, then the measured values
are likely to be smaller than the actual value because approximately 15% of the COD is lost
during the centrifugal drying process. Nevertheless, the slope of the regression equation with
no intercept (from 15.6-16.3 kJ/g COD) still indicates a greater energy per gram of COD than
is present in a selection of organic compounds commonly found in municipal wastewaters
(Table 4-4) (Huang et al., 2010). The rest of the content of this Section 4.1.3 will discuss the
cause of this difference.
Table 4-4. Common organic compounds and their calculated energy values
Compoundsa Type Formula ∆H/gCODb

Glutamic Proteins C5H9NO4 13.4
Aspartic C4H7NO4 13.3
Glucose Sugars C6H12O6 14.6
Xylose C5H10O5 14.7
Volatile Fatty
Acetic acid CH3COOH 13.6
Acids
Butyric Acid CH3CH2CH2COOH 13.6
Humic
Fulvics C33H32O19 15.7
substances
Humics C34H34O16N2 11.6-14.5
Cellulose Others (C6H12O6)n 14.6
Lignin (C9H10O2)n, (C10H12O3)n or (C11H14O4)n 14.1
a
A selection of different compound types selected, but all have been shown to be present at high
concentration in municipal wastewater by Huang et al. (2010).
b
∆H/gCOD is calculated from deoxygenation enthalpy values presented by Sato (1990), except for
Fulvics which is based on data in Reddy et al. (1989). Values for energy content of additional organic
compounds can be found in Heidrich et al. (2010).
77
In principle, the kJ/g COD can be calculated via the Equation 4-1.
Equation 4-1
𝐸𝑤𝑤
𝑘𝐽⁄𝑔 𝐶𝑂𝐷 =
𝐶𝑂𝐷
Where Eww (kJ/L) is the chemical energy of a wastewater and COD (g/L) is the COD
concentration of that wastewater.
There are three main issues that might compromise the accuracy of the determination of the
kJ/g COD value:
A. Whether the bomb calorimeter is able to combust all the energy-containing substances
B. Whether the dichromate method of COD analysis is able to oxidize all the substances
that contribute energy and
C. How much of the substances that contribute to energy is lost during the drying process.
The quite a few common organic compounds listed in Table 4-3 have been measured for their
calorific value via bomb calorimetry, for instance, Shizas and Bagley (2004) measured glucose,
Evans and Skinner (1959) measured the acetic acid, Ribeiro da Silva et al. (2013) measured
cellulose, Colbert et al. (1981) measured cellulose, Tsuzuki and Hunt (1957) and (Tsuzuki and
Hunt, 1957) measured aspartic acid and glutamic acid. In those studies, the measured caloric
values of the tested substances well match those substances’ standard enthalpy of combustion
since the difference between the two is less than 5% (Shizas and Bagley, 2004, Colbert et al.,
1981, Evans and Skinner, 1959, Ribeiro da Silva et al., 2013, Wu Yang et al., 1999). This
indicates the bomb calorimeter is able to fully combust those substances in the wastewater and
deliver a reliable calorific value.
Regarding the second issue that may compromise accuracy (B), in this research the COD was
tested using the dichromate method, in which the contaminants are oxidized in a solution rich
in the dichromate ion (Cr2O72-) (APHA, 2012). During the reaction, the oxidant Cr2O72- is
reduced to Cr3+, which results in a color change in the solution (APHA, 2012). Subsequently,
the quantification of the extent of the color change indicates the COD concentration of a sample
(APHA, 2012). But the dichromate is unable to fully oxidize every oxidizable substance, and
yet some of these oxidizable substances can contribute chemical energy (APHA, 2012, Li et
al., 2012, Boyles, 1991). In this sense, the COD measured by the dichromate method may be
78
lower than the theoretical COD of a sample completely oxidized via combustion. This could
therefore result in an erroneously high value of kJ/g COD.
The substances that contain energy but are resistant to Cr2O72- oxidation include benzene,
toluene, pyridine, ammonia and its derivatives (considered as the TKN). (Li et al., 2012,
Scherson and Criddle, 2014, Boyles, 1991). The contribution of these substances to the total
Eww varies. Although the calorific value of benzene, toluene and pyridine can be up to two
times greater than that of TKN, their combined contribution is actually less than 1% because
of their low concentrations in wastewater (Huang et al., 2010, Li et al., 2012, Thornton et al.,
2001, Zalat and Elsayed, 2013). Consequently their contribution to energy is less than one tenth
of the energy contained in TKN (shown in Table 4-5). Scherson and Criddle (2014) suggested
that approximately 15% of chemical energy in raw wastewater is from TKN. Hence, this
research considers that the energy contributed from the unmeasurable COD is mainly from
TKN. Total wastewater energy content is considered to come mainly from TKN and the
measurable COD.
Table 4-5 Chemical energy contributed by commonly found substances unaccounted for in the COD
measurement
Concentration in raw Calorific Chemical Proportion to
a
wastewater, mg/L value, kJ/g b
Energy, kJ the total Eww, %c
Benzene 0.2 41.8 0.01 0.1%
Toluene 0.6 42.4 0.03 0.3%
Pyridine 1.0 33.5 0.03 0.4%
+ c
NH4 -N 27.2 22.5 0.61 7.0%
c
TKN 40.77 22.5 0.92 10.6%
a
For benzene and toluene, the listed value is the reported in Thornton et al. (2001). For Pyridine, the listed value
is based on data reported in Zalat and Elsayed (2013). For N-NH4+ and TKN, the listed values are the mean
concentration of the composite raw.
b
The caloric value is calculated by dividing the enthalpy of combustion by the molar weight (NIH, 2018b, NIH,
2018a, NIST, 2018b, NIST, 2018a).
c
The proportion is calculated by diving the chemical energy by the mean Eww of the composite raw (listed in Table
4-2).
Ammonia is volatile and is likely to be evaporated during the drying process along with other
volatile compounds that contribute to COD (Maurer and Müller, 2012, Huett, 1997). Hence, its
contribution to the subsequent energy measurement will be cancelled. But the nitrogen
containing protein may still stay in the dried residue after evaporation (Singh, 2007). The exact
79
amount of NH4+-N and TKN that remained in the dried residue was not determined as part of
this research. Hence, it was not possible to calculate the exact energy contributed by the NH4+-
N and TKN compared to that contributed by the measurable COD. In this sense, the slope no
longer indicates how many kJ of chemical energy is contained in a gram of COD but just the
empirical mathematical relationship between COD and Eww. Based on the regression
determined in the Figure 4-1, the Eww of wastewater can be estimated via Equation 4-2 which
is the regression equation from plotting Eww of All Samples against COD .
Equation 4-2
𝐸𝑤𝑤 = 𝐶𝑂𝐷 × 0.0158 kJ/mg

where COD is the chemical oxygen demand of a wastewater and is in unit of mg/L.
Fundamentally, to measure the chemical energy content of wastewater it is necessary is

calculate the product of the concentration and heating value of each energy-containing
chemical within the wastewater, and then sum the products for all such chemicals in the
wastewater. This would be particularly useful for the purposes of energy audits and decision-
making regarding selection and implementation of energy recovery technologies. Taking a
lignin-rich wastewater as an example, mechanical separation and subsequent recovery would
be a better option than anaerobic treatment which is less efficient at breaking down lignin
(Stoica et al., 2009). However, wastewater is a mixed waste and the number of individual
compounds it contains could be measurable in hundreds or thousands (Huang et al., 2010). It
is likely that the composition of the energy-containing chemicals are site-specific (Heidrich et
al., 2010). Therefore, the work required for understanding the concentration and heating value
of each energy-containing chemical would be substantial and does not tend to be practical, or
cost-effective, for a general energy audit of numerous WWTPs or for initial technical feasibility
assessment of a particular technology. But the regression equations derived as part of this
research, which are determined from various types of samples collected from four WWTPs of
varied population equivalents and locations, indicates that a reliable determination of energy
content can be derived from the COD alone, which is quickly and cost-effectively determined.
4.1.4 The impact from diurnal variation to the chemical energy in wastewater
The diurnal variation of influent wastewater quality has a direct bearing on the relationship
between Eww and COD. As shown in Figure 4-1 (F), the inclusion of spot raw sample data in
80
the regression does not result in any substantial change in the relationship (i.e the slope of the
regression line) between Eww and COD. However, Figure 4-1(E) shows that the regression plots
using spot raw water sample data only has more scatter than those regressions using composite
sample data. This research did not perform a detailed study of the variance of organic load of
the spot raw wastewater samples, but the ratio between pCOD and COD of each spot and
composite raw sample have been analyzed. Although there is no statistical difference between
the mean value of the pCOD:COD ratio of all spot raw samples compared to all the composite
raw samples (p=0.308), there is significant difference in the variance (Levene test, p>0.05)
(Minitab Inc., 2017c). Individual value plots of two groups of data is shown Figure 4-2. Diurnal
variability is considered to be the cause of the greater scatter in the regressions involving spot
raw analyses only, but further investigation of the exact reasons for this variation of chemical
composition is recommended.
100.0%
90.0%
80.0%
pCOD:COD
70.0%
60.0%
50.0%
40.0%
30.0%
Composite Raw Samples Spot Raw Samples
Figure 4-2 Individual value plot of pCOD:COD of Composite Raw and Spot Raw
Apart from influencing the determination of relationship between Eww and COD, the diurnal
variation also impacts the chemical loading to the WWTPs. The hourly chemical energy
loading to WWTP A and WWTP C were investigated as part of this research. These plants
have population equivalents of 40,000 and 1,000,000, respectively. The hourly flow, COD and
energy loading to WWTPs A and C are shown in Figure 4-3.
81
For those hourly samples for which Eww was not directly measured, it was estimated via
Equation 4-2 i.e. by multiplying the hourly flow by COD (mg/L) and by 0.0158 kJ/mg COD
(shown in Table 4-3, row 6)
(1) WwTP A
2000 1100
Flow, m3/hr or COD, mg/L

1800
1600 900
Energy, kWh/hr
1400 700
1200
1000 500
800
600 300
400 100
200
0 -100
13:00
09:00
10:00
11:00
12:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
00:00
01:00
02:00
03:00
04:00
05:00
06:00
07:00
08:00
Energy, kWh/hr Flow, m3/hr COD, mg/L
(2) WwTP C
45000 1400
Flow, 10 m3/hr or COD, mg/L

40000 1200
35000
Energy, kWh/h
1000
30000
25000 800
20000 600
15000
400
10000
5000 200
0 0
10:30
11:30
12:30
13:30
14:30
15:30
16:30
17:30
18:30
19:30
20:30
21:30
22:30
23:30
00:30
01:30
02:30
03:30
04:30
05:30
06:30
07:30
08:30
09:30
Energy, kWh/hr Flow, 10 m3/hr COD, mg/L
Figure 4-3 The hourly flow, COD and energy loading to the WWTP A and C
The pattern of hourly energy loading to the WWTPs is site specific. In terms of the hourly
flowrate, variation is similar for the two sites; using the ratio of maximum flow to minimum
flow as a metric, 2.0 for WWTP A and 2.4 for WWTP C. But WWTP A has a greater diurnal
variation of COD concentration, as its ration maximum to minimum COD is 7.8, while for
WWTP C it is only 2.3. This phenomenon matches the reported finding that the diurnal
variation of COD can be large in small treatment works (Raboni et al., 2013). For the hourly
82
energy loading, the maximum to minimum ratio is 15.2 for WWTP A and 5.7 for WWTP C.
Hence, at smaller sites it can tentatively be concluded that there will be greater variability in
energy load.
There is an increasing need to recover energy from wastewater (Scherson and Criddle, 2014,
Wan et al., 2016). COD is not only an indicator of the Eww of municipal wastewater but also
one of the decisive factors determining the feasibility of an energy recovering technology. Stoll
et al. (2018) report that microbial fuel cells (MFCs) will only be energy neutral or energy
positive if the inflowing COD concentration is high enough and if there is high enough removal
of COD. The diurnal fluctuation of COD is therefore likely to affect the energy balance of the
MFC (Stoll et al., 2018). To recover energy effectively a successive technology should be able
to convert the COD into an energy product efficiently and be able to adapt to the diurnal
variation of both loading and types of COD. Therefore, more work should be done to
understand the diurnal fluctuations of COD and energy loading and the variation of types of
COD or energy material flows to the WWTP.
4.2 Chemical energy of sludge
4.2.1 Sludge Characteristic
Summary statistics for TS%, VS/TS%, TOC/TS%, TN/TS%, TP/TS%, and the chemical energy
content of the sludge of all four study sites are shown in Table 4-6.
To begin with the primary sludge and Surplus Activated Sludge (SAS) as secondary sludge
generated from 4 WWTPs are compared. Although the primary sludge produced in WWTP A
and WWTP D both contain SAS which gets recirculated from the final clarifiers to the primary
settlement for co-settling, there is no statistical difference (two Sample T test, p>0.05) in
chemical energy content of dried sludge (Eds, kJ/g) of primary sludge and SAS between the
four WWTPs.
WWTP B and WWTP C are both district sludge thickening and dewatering centers and have
mechanical sludge thickening and dewatering treatment comprising drum thickening and
83
centrifugal dewatering. WWTP C also uses thermal treatment via CAMBI thermal hydrolysis,
and biological sludge processing using anaerobic digestion. The Eds of the sludge before and
after each sludge treatment process was studied in this research. There is no significant
difference (p>0.05) in Eds of the sludges before and after the mechanical and thermal-chemical
treatment. But the Eds of the sludge fed to the digester is significantly higher than the energy
content of digested sludge. The same applies to VS/TS% and TOC/TS%. Therefore, anaerobic
digestion appears to be lowering the Eds of the sludge. The mean value of the energy content
of the dried and wet sludges and TS%, VS/TS%, TOC/TS%, TN/TS% and TP/TS% of the
sludge collected are shown in Table 4-6.
84
Table 4-6 The mean value (± standard deviation) of TS%, VS/TS%, TOC/TS%, TN/TS%, TP/TS% and chemical energy content of the dried sludge and wet sludge
Type n TS% VS/TS% TOC/TS% TN/TS% TP/TS% Eds, kJ/g* Es, kJ/kg**
Primary Sludge 20 3.3±0.7 % 82.6±3.5 % 42.6±3.1 % 3.2±0.6 % 1.2±0.3 % 16.8±1.2 563.8±124.8
Thickened Primary Sludge 5 5.8±1.0 % 81.5±2.5 % 40.8±1.7 % 2.9±0.5 % 1.0±0.2 % 16.5±0.9 950.2±135.6
SAS 20 0.5±0.2 % 78.2±5.1 % 31.6±2.9 % 6.2±0.5 % 2.2±0.8 % 16.4±1.1 72.9±31.5
Thickened SAS 10 3.9±1.2 % 83.2±1.4 % 38.5±2.6 % 6.7±0.5 % 2.7±0.6 % 18.1±0.5 714.2±226.4
Stored Sludge 10 3.6±0.7% 78.3±6.1 % 38.9±3.3 % 4.6±0.8 % 1.6±0.1 % 17.3±1.4 620.6±120.9
Dewatered Stored Sludge 10 23.5±2.0 % 82.9±1.6 % 41.3±2.7 % 3.7±0.4 % 1.3±0.2 % 18.2±0.5 4266.0±367.0
THP Feed 5 15.7±1.3 % 82.0±1.0 % 39.9±0.7 % 3.2±1.1 % 1.8±1.1 % 18.1±0.9 2861.0±322.0
Digester Feed 5 10.3±0.9 % 82.2±1.2 % 42.2±1.3 % 4.1±0.5 % 1.1±0.1 % 18.6±0.8 1921.0±212.4
Digested Sludge 5 5.3±0.4 % 67.8±1.6 % 35.5±1.1 % 8.6±1.4 % 2.1±0.3 % 15.6±0.7 824.0±90.1
Dewatered Digested Sludge 5 28.5±3.4 % 70.1±7.7 % 34.8±1.1 % 5.2±0.2 % 1.8±0.0 % 15.2±0.3 4322.0±485.0
*: Eds is the acronym of chemical energy content of dried sludge

**: Es is the acronym of the chemical energy content of the wet sludge
88
4.2.2 The relationship between solids and chemical energy content of sludge
As with the study of Eww, surrogates that could be used to reliably indicate sludge energy
content were evaluated. The weight to weight (w/w) concentration to the total solid content,
such as VS/TS%, TOC/TS%, and the weight to volume (w/v) concentration, such as the TS s,
VSs, are analyzed for their relationship with the chemical energy content of the dried sludge
(Eds, kJ/g) and with the chemical energy content of wet sludge (Es, kJ/L).
In all cases the w/w concentration to the total solid content (i.e. VS/TS%, TN/TS%, etc.) had a
weak (<0.50) correlation with both Eds and Ews. All the w/v concentrations to the sludge (in
unit mg/L) are found with a weak correlation with the Eds, but with strong correlation (>0.89,
p<0.05) with the Es. All the w/v concentrations are found highly correlated with each other
(Spearman Rho, >0.900, p<0.05). Therefore, the w/v concentrations are selected to proceed the
best subset study. The results of the Best Subset study are shown in Table 4-7.
Table 4-7 Best regression for estimating wastewater energy
No. of Predictor r2 r2adj Parameter Required
1 99.4 99.4 TOCs
1 99.4 99.4 VSs
1 98.5 98.5 TSs
1 83.4 83.2 TNs
1 80.8 80.6 TPs
2 99.6 99.6 VSs, TOCs
3 99.6 99.6 TSs, VSs, TOCs
4 99.6 99.6 TSs, VSs, TOCs, TNs, TPs
5 99.6 99.6 TSs, VSs, TOCs, TNs
The single variable regressions determined by each w/v concentration to sludge energy has
strong linear relationship (r2>0.80, p<0.05) with Es. The r2 of the best three single variable
regressions, determined by TOC%, VS% or TS%, are nearly identical to the r2 (99.6%) of the
best model. Therefore, the single variable regression is already good enough to predict the Es.
The plot, regression line, confidence interval and prediction interval of the regression
determined by TOCs, VSs or TSs are shown in Figure 4-4. All the regression shown in Figure
4-4 considered intercept in the regression equation. The determined regression equations with
and without intercept are shown in Table 4-8. It is worth noting that the intercept is found not
statistically significant (p>0.05) in all three cases.
89
(A)
(B) (C)
Figure 4-4 Plots of Es against TSs, VSs and TOCs (A) Es against TSs, (B) Es against VSs, and (C) Es against TOCs
Table 4-8 Regression equations of regressions between Es and TSs, VSs and TOCs shown in Figure 4-4
Regression Equation with Regression Equation without
n
intercept* intercept*
Figure 4-4 (A) 87 Es=23.8+0.0170 TSs (r2=98.5%) Es= 0.0171 TSs (r2=99.2%)
Figure 4-4 (B) 87 Es=-10.5+0.0220 VSs (r2=99.4%) Es= 0.0219 VSs (r2=99.7%)
Figure 4-4 (C) 87 Es=2.5+0.0437 TOCs (r2=99.4%) Es= 0.0438 TOCs (r2=99.7%)
*: Es is in unit of kJ/L. TS, VS and TOC are in unit of mg/L.
Similar to the study of the chemical energy of wastewater, this research aimed to identify a
reliable empirical method to quickly estimate the energy content or loading of a sludge using
routinely analysed variables such as TS% and VS/TS%, or TSs and VSs that are interpreted
from TS% and VS/TS%. Figure 4-4 suggests that the Es can be estimated from TSs, VSs and
TOCs. Using the TSs for estimation would be the simplest and quickest because both the VSs
and TOCs is measured on top of the measurement of TSs.
In regression equation shown in Table 4-8, Es is in unit of kJ/L and TSs is in unit of g/L, thus
the slope should be in unit of kJ/g TS. The slope 17.1 kJ/g TSs (in Figure 4-4 (A)) is nearly
90
identical to 17.11±1.33 kJ/g which is the mean value of Eds of all the sludge sample analyzed
in this study. It is likely that the calorific value of the dried sludge, the Eds, determines the slope.
However, Eds values from literatures are substantially varied, as shown in Table 4-9. The mean
Eds values determined by Shizas and Bagley (2004) and Schaum et al. (2016) are significantly
smaller (Two Sample T Test, p<0.05) than the mean Eds found in this research. Thus, there may
be risks of errors associated with using 17.10 kJ/g TSs to evaluate the Es.
Table 4-9 Energy related characteristic of sludge reported in literature

Primary Sludge SAS Digested Sludge
Eds, VS/TS%, kJ/g Eds, VS/TS%, kJ/g Eds, VS/TS%,
kJ/g VS
kJ/g % VS kJ/g % VS kJ/g %
Shizas and
Bagley 15.90 67.00 23.73 12.40 60.00 20.66 12.70 51.00 24.90
(2004)
Schaum et
16.40 78.00 21.02 14.30 69.00 20.72 11.10 59.00 18.81
al. (2016)
By dividing the Eds with VS/TS%, kJ/g VS is obtained and illustrates how many kJ of energy
is contained in a gram of VS. There is no significant difference (Two Sample T test, p>0.05)
between the mean kJ/g VS found in this study and that found by Shizas and Bagley (2004) and
Schaum et al. (2016). Moreover, the mean value of kJ/g VS of this study is 21.48 kJ/g. It is also
nearly identical to 21.89 kJ/g VS which is the slope of the regression equation determined by
VSs (shown in Figure 4-4 (B)). Therefore, Es can be estimated from VSs using a value of 21.9
kJ/g VS and Equation 4-3.
Equation 4-3
𝐸𝑠 = 𝑉𝑆𝑠 × 21.9 = 𝑇𝑆𝑠 × 𝑉𝑆/𝑇𝑆% × 21.9
The major components in the sludge are the carbohydrate-lignin organic compounds, such as
cellulose, hemicellulose and lignin (Ogilvie, 1998). Considering they can be fully combusted,
their caloric values are the same as their kJ/g VS values, which are 16.5 kJ/g for cellulose, 13.9
kJ/g for hemicellulose, and 20.7 kJ/g for lignin (Kim et al., 2017). Although they are lower
than the value of 21.48 kJ/g VS, sludges also contains substances with higher caloric values
such as the grease (Ogilvie, 1998). The grease can consist of Myristic Acid, Palmitic Acid,
Stearic Acid, Oleic Acid and Linoleic Acid (Dong, 2015). Their calorific values range from 37
91
to 40 kJ/g (Demirbas, 2016). Therefore, estimated Es via Equation 4-4 appears to be a
reasonable approach.
Current practice is mainly to recover energy from wastewater via anaerobic digestion of sludge.
Maximizing the sludge production from wastewater treatment is one of the measures to
improve the energy recovery and energy efficiency of a WWTP (Wan et al., 2016). Via
Equation 4-1 and 4-4, researchers and operators can quickly estimate the energy flux of a
treatment process or a treatment work using the routinely measured parameters of COD and
VS/TS%. This will help identifying potential opportunities for maximizing sludge production
and discovering potential opportunities for energy recovery.
4.2.3 The relationship between COD and chemical energy content of sludge
Although the ES can be estimated via VSs, there is still difficulty in applying modelling to
construct the chemical energy balance, since VSs is not able to predict in the modelling software
GPS-X. Thus, a relationship between the VSs and the COD in the sludge has to be defined.
By definition, the COD of the sludge is contributed from the VS (as shown in Figure 4-5)
(Foladori et al., 2010). The VSs of the sludge, which consists of both pCOD and sCOD, is
originally from the wastewater. If in the wastewater there is a strong relationship between COD
and VS, this relationship will possibly pass on to the sludge. Subsequently, the relationship of
COD and Es can be determined.
Figure 4-5 Physico-chemical fractionation of total COD in the sludge (Foladori et al., 2010)
92
The composite samples are used for this investigation since both COD and VS are measured
on them. The statistical analysis suggests that there is a strong linear relationship (n=61,
r2=92.13%, p<0.05) between the two as shown in Figure 4-6.
Figure 4-6 The relationship between COD and VS in wastewater
In the wastewater, COD (mg/L) can be estimated via VS (mg/L) via Equation 4-4
Equation 4-4
COD = −66.5 + 1.3526 VS
The study then applied Equation 4-4 to estimate the sludge COD. The estimated sludge COD
is found having a strong linear relationship (r2=99.66, p<0.05) with the sludge energy Es (kJ/L),
as shown in Figure 4-7. The regression equation determined are Es= -9.5+0.0162 COD or Es=
0.0162 COD. The slope 0.0162 falls the range of 0.0156 to 0.0171 which are the slope
determined for the COD and Eww via different sample groups (shown in Table 4-3). The
relationship between the energy content (kJ/L) and COD (mg/L) in both wastewater and sludge
are nearly identical. The implication of it is that the chemical energy balance is very likely to
be the same as the COD mass balance in the wastewater treatment and in the sludge treatment
in this four studied WWTPs. Therefore, this study will use Equation 4-2 to estimate the
wastewater and sludge energy and will use the COD mass balance to express the chemical
energy balance.
93
Figure 4-7 The relationship between energy content (kJ/g) and COD in sludge
4.3 Chapter Summary
This research successfully developed a new wastewater drying method that is at least as good
as the proven method of freeze drying for retaining the energy-containing material but is a far
less time-consuming method.
Subsequently, this research conducted statistical analysis (correlation and Best Subset of
regression) to investigate the relationship between Eww and commonly analysed wastewater
variables. COD was found to be the best surrogate of energy content of wastewater. It has a
ubiquitously (across the four WWTPs) strong linear relationship (77-93%, p<0.05, excluding
the spot raw) with Eww. Therefore, an empirical equation is proposed for estimating the
wastewater energy from COD.
A similar statistical study was also conducted for understanding the relationship between Es
and the other commonly analysed sludge variables. Strong linear relationships (>85%, p<0.05)
were found between the Es and TSs, VSs or TOCs. Empirical equations are proposed for
estimating the sludge chemical energy from the TSs and VSs. Moreover, via assuming the
wastewater and sludge share the same relationship between COD and VS, the sludge COD was
estimated. Subsequently, the relationship between chemical energy and COD in the sludge is
found similar in the wastewater. This implies that chemical energy balance of a WWTP shall
follows the COD mass balance.
94
The research aims to develop a quick but reliable method of estimating wastewater energy
content using routinely analyzed variables has been fulfilled. As the energy of wastewater and
sludge can be estimated much more quickly as a consequence, an energy audit of a treatment
process or a treatment works can now be easily conducted. This will help researchers and
operators to better monitor the energy recovery performance and discover potential
opportunities for improvement in energy recovery.
Wastewater is a mixture and hence its energy source is from multiple types or groups of
chemicals. This research did not investigate the breakdown of the energy contribution from
different groups of chemicals as it was not a primary aim of the research. Nevertheless, that is
an area recommended for future research because it may help target energy recovery
technologies to specific energy containing materials that are not recovered using current
methods.
95
Chapter 5 COD Mass Balance and Energy Balance of Wastewater
Treatment Plants
This chapter first presents the calibration and validation of the wastewater and sludge
modelling of four WWTPs (Section 5.1). The COD and energy balances of each process are
shown in Section 5.2. Based on these balances, in Section 5.3 there is a discussion of the
potential for energy recovery from wastewater and sludge. Section 5.4 builds four case studies
to examine how energy recovery approaches might impact on the electricity balance of a
conceptual WWTP.
5.1 Model calibration and validation
5.1.1 Wastewater and sludge treatment models
5.1.1.1 Wastewater treatment models

Wastewater treatment models were built for WWTP A, WWTP B, WWTP C, and WWTP D.
Screen shots of the models are shown in Figure 5-1 to 5-4. In each case the process, the flow
diagram represents the actual treatment units at each WWTP, with the exception of the virtual
units added for sCOD and PO43--P removal, the virtual unit added for controlling the biological
reaction in the settlers, and the simplication of the trickling filter process of WWTP A
(explained in Section 3.4.1.2.2).
It is worth noting that in all four WWTPs there is an sCOD decrease of 15 mg/L during primary
settlment. This is considered to be due to flocculation and subsequent settling of the sCOD
(Gernaey et al., 2001). Thus, a virtual removal unit is added to each wastewater treatment
model to ensure this sCOD decrease is reflected in the model.
In WWTP B there is also a decrease of >1.5 mg/L PO43--P during primary treatment. This is
thought to be due to sorption of phosphate by iron and subsequent settlement. The elevated iron
concentration is thought to be due to mining-related waters entering the WWTP (L Wilkinson,
personal communication, 29 October 2018). A virtual PO43--P unit is added to the WWTP B
model, prior to its primary settlement, to simulate this.
96
In the case of WWTPs A and D, biological sludges is recirculated back to the primary
settlement. Moreover, their activated sludge process has nitrification and potential
denitrification may occurred in final clarification. Thus, the primary settlements and final
clarifications of WWTPs A and D are biological reactive. In their model, virtual activated
sludge tanks (“Plug Flow Tank” unit) are added to control the biological reaction.
Figure 5-1 The wastewater treatment model of WWTP A
Figure 5-2 The wastewater treatment model of WWTP B
97
Figure 5-3 The wastewater treatment model of WWTP C
Figure 5-4 The wastewater treatment model of WWTP D
The influent water quality data, and other input variables, used for the modelling of each of the
WWTPs are listed in Table 5-1.
98
Table 5-1 The influent water quality, and other input variables, used for the modelling of each WWTP
WWTP WWTP WWTP WWTP
Parameter
A B C D
Flow, m3/d 9052 49624 278954 37312
COD, mg/L 1061.4 727.1 665.0 336.6
TKN, mg/L 50.9 56.9 55.50 31.8
TP, mg/L 8.5 9.1 10.8 4.0
NH4--N, mg/L 30.1 37.7 41.8 19.9
Nitrite-N, mg/L 0.03 0.0 0.00 0.02
Nitrate-N, mg/L 0.23 0.0 0.00 0.28
PO43--P, mg/L 3.8 7.1 6.65 1.50
VSS/TSS 0.83 0.85 0.81 0.78
Soluble inert fraction of total COD (frsi) 0.050 0.050 0.050 0.050
Readily biodegradable fraction of total COD (frss) 0.200 0.200 0.200 0.200
Particulate inert fraction of total COD (frxi) 0.130 0.130 0.130 0.130
Colloidal fraction of slowly biodegradable COD
0.014 0.148 0.075 0.161
(frscol)
NH4:sTKN 0.800 0.900 0.800 0.800

N content of soluble inert material (iNSI)
0.100 0.050 0.100 0.100
, gN/gCOD
N content of inert particulate inert material (iNXI)
0.030 0.030 0.030 0.030
, gN/gCOD
P content of soluble inert material (iPSI)

0.005 0.005 0.005 0.005
, gN/gCOD
P content of inert particulate inert material (iPXI)
0.005 0.005 0.005 0.005
, gN/gCOD
TOC/COD 0.293 0.323 0.311 0.341
5.1.1.2 Sludge treatment models

Although both WWTPs B and C have sludge thickening and dewatering, because of lack of
data of the sludge thickening and dewatering of WWTP B, only the sludge dewatering of
WWTP C has been modelled. The sludge thickening and dewatering prior to digestion is built
on the wastewater treatment model of WWTP C (Figure 5-3). It is worth noting that 300 m3 of
pure water is virtually added to the raw sludge tank in order to make up the flow of the raw
99
sludge to approximately 2,300 m3 that matches the historical data. The anaerobic digestion
model is built separately. The sludge treatment models are shown in Figure 5-5 and 5-6.
Figure 5-5 The model of WWTP C’s wastewater treatment and sludge thickening and dewatering prior to
anaerobic digestion
Figure 5-6 The model of WWTP C’s anaerobic digestion and the subsequent sludge dewatering
5.1.2 Calibration and validation of the wastewater treatment models
5.1.2.1 Calibration of the wastewater treatment models

In all four WWTPs the aerobic heterotrophic yield on soluble substrate always ranks in the top
three. That is because it is influential to variety of the targeted variables and the the most
important ones found in this study are the flow of SAS, sCOD, PO43--P concentration. This
100
phenomenon has been identified in previous studies (Cosenza et al., 2014, Sin et al., 2011,
Mannina et al., 2011).
The settling parameters of the primary settlers and final clarification are classified as influential.
That is because the settling parameters of the primary settler control quantity of slowly
biodegradable flows to the activated sludge process and the parameters of the final clarification
control the concentration of the pCOD.
Maximum growth rate for ammonia oxidizer and ammonia oxidizer aerobic decay rate were
found to be influential to ammonium content of the final effluent (Sin et al., 2011, Cosenza et
al., 2014). The modelling of WWTP A, B and D agrees this phenomenon, but not for WWTP
C. The reason is that the HRT of WWTP C’s activated sludge process is too short to allow
ammonium oxidation to happen. Because nitrification occurs in WWTPs A and D, some of the
nitrification and denitrification related parameters are found to be influential in these two
WWTPs.
However, the parameters controlling phosphate removal by phosphate accumulating organism

(PAOs)re rarely ranked in the top 20 or 30. That is because none of the WWTPs have an
anaerobic or anoxic section in their activated sludge process (Tchobanoglous et al., 2003).
Phosphate removal by PAO therefore rarely happens.
The most important parameters of influent fraction, kinetic and stoichiometric that influence
the modelling results, and are therefore subject to adjustment in the model for each WWTP, are
listed in the Tables D-1 to D-4 in Appendix D. The table also presents the δjmsqr (that indicates
the influence of the parameter to the entire model), the adjusted value, the order of adjustment,
and the category each adjusted parameter belongs to.
101
5.1.2.2 Validation of wastewater treatment models
For the wastewater treatment modelling, the ARD between the predicted results and the
observed results of WWTPs are:
• WWTP A: 28.1%
• WWTP B: 11.3 %
• WWTP C: 3.9%
• WWTP B: 13.8%
The high ARD of WWTP A is due to the predicted NH4+-N concentration (0.88 mg/L) is 338%
greater than the observed NH4+-N concentration (0.22 mg/L). However, Rieger et al. (2012)
suggested that it is allowable to have a 1 mg/L difference between the predicted result and the
observed result when the concentration is low. In this case, the observed and predicted NH4+-
N concentration is lower than 1 mg/L and the difference between the two is only 0.68 mg/L
which is smaller than the allowable difference. If the NH4+-N concentration is excluded, the
ARD of WWTP A becomes 10.8%. Then, the ARD of the modelling of all four WWTPs are
then no greater than 15%. As shown in Figure 5-7, the observed results well match the model
predictions.
Figure 5-7 Comparison between the observed results and predicted results of the modelling of four WWTPs
(A)WWTP A,(B) WWTP B, (C)WWTP C and (D) WWTP D
102
Hence, the model constructed for the wastewater treatment plants is considered representative
of the actual performance and the material mass balance of the WWTPs. The predicted flowrate
and concentrations are therefore used in the mass and energy balance calculations for the
WWTPs in this chapter and in Chapter 6.
The observed and predicted result of the wastewater modelling are shown in Table E-1 in
Appendix E.
5.1.3 Calibration and validation of the sludge treatment models
5.1.3.1 Calibration of the sludge treatment models prior to anaerobic digestion

In the sludge storage, thickening and dewatering prior to the digestion, the following factors
are found to be most influential on the biological reactions of the sludge storage tank: yield of
fermentative biomass, aerobic heterotrophic decay rate, anaerobic reduction factor for decay
rate (in Heterotrophic Biomass section of the biological model), hydrolysis rate constant for xs,
saturation coefficient for particulate COD, anaerobic reduction factor for decay rate (in
Heterotrophic Biomass section of the biological model). These factors are influential because
the majority of the reactive biomass within the sludge is heterotrophic biomass. They affect the
solubilization of the particulate substrate and ultimately impact the soluble component of
sCOD, ammonium and phosphate of the filtrate and centrate.
For the thickening and dewatering, the parameter fraction of influent flow of the sludge
thickening and dewatering is always found to be influential as it affects the TS content of the
thickened or dewatered sludge. The removal efficiency of slowly biodegradable substrate, inert
particulate material and heterotrophic biomass (particularly in the SAS thickening) are found
to be influential on the modelling results, especially to the pCOD content in the filtrate and
centrate. The removal efficiency of nitrogen in slowly biodegradable substrate and phosphorus
in slowly biodegradable substrate mainly impacts on the TP and TN content of the thickened
and dewatered sludge and the TKN content of the filtrate and centrate.
The parameters of kinetic and stoichiometric parameters that influence the biological modelling
results of the sludge storage and sludge thickening and dewatering processes prior to digestion,
103
and are therefore subject to adjustment in the model for each WWTP, are listed in the Tables
D-5 to D-7 in Appendix D. The table also presents the adjusted value and the category those
adjusted parameter belongs to.
5.1.3.2 Calibration of the sludge treatment models of anaerobic digestion and the subsequent
sludge dewatering
In the modelling of the anaerobic digestion and its subsequent dewatering, the particulate inert
fraction of total COD and yield of fermentative biomass is found to be influential on the extent
of the digestion and the volume of biogas produced. The C content of inert particulate material
and of the slowly biodegradable substrate are influential to the ratio of CO2 and CH4 in the
biogas (Hydromantis, 2015).
For the digested sludge centrate, N and P content of the inert particulate material influences the
amount of N and P in the particulate biodegradable substrate which can be digested and release
ammonium and phosphate. Therefore, these two fractions are influential to the ammonium and
phosphate content of the centrate. The fraction of inert COD during slow biodegradable organic
hydrolysis affects the production of inert sCOD and influences the sCOD concentration in the
centrate.
Like sludge dewatering, the fraction of influent flow and removal efficiency of particulate inert
material of the sludge dewatering unit are influential the solids content of the dewatered sludge
and the pCOD concentration of the filtrate and centrate. Unlike the sludge thickening and
dewatering prior to digestion, the wet digested sludge is rich in unbiodegradable cells and
biomass related to the digestion process but has limited heterotrophic biomass or slowly
biodegradable substrate (since it has been digested), as shown in Table 5-2. Therefore, the
removal efficiency of the unbiodegradable cell product, fermentative biomass, acetogenic
biomass, acetolactic methanogenic biomass, and hydrogenotrophic methanogenic biomass also
influences the particulate content of filtrate and centrate.
104
Table 5-2 The predicted different type of CODs of the feed sludge to different sludge thickening and dewatering
Feed Sludge Feed Feed
Feed
of primary Sludge of Sludge of
Sludge of
sludge raw digested
SAS drum
drum sludge sludge
thickening
thickening centrifuge centrifuge
Total COD, mg/L 37490.0 6948.0 46870.0 40830.0
Particulate inert material, mg/COD 7694.0 1830.0 12720.0 23810.0
Unbiodegradable cell product, mg/COD 669.1 81.5 1408.0 3072.0
Slowly biodegradable substrate, mg/COD 26210.0 1306.0 23060.0 28.3
Heterotrophic biomass, mg/COD 255.2 3683.0 3443.0 0.0
Fermenting biomass, mg/COD 0.0 0.0 0.0 7406.0
Acetongenic biomass, mg/COD 0.0 0.0 37.9 14.7
Acetoclastic methanogenic biomass, mg/COD 0.0 0.0 0.0 1433.0
Hydrogenotrophic methanogenic biomass, mg/COD 0.0 0.0 139.1 1073.0
The parameter of influent fraction (of digester feed sludge), kinetic and stoichiometric
parameters that influence the biological modelling results of the sludge storage, sludge
thickening and dewatering processes and the anaerobic digestion, and are therefore subject to
adjustment in the model for each WWTP, are listed in the Tables D-8 in Appendix D. The table
also presents the adjusted value and the category those adjusted parameter belongs to.
5.1.3.3 Validation of the sludge treatment models

For the sludge treatment modelling, the ARD between the predicted result and the observed
results are also no higher than the allowable value of 15.0%. The ARDs of the various treatment
stages are as follows:
• primary sludge thickening: 7.4%
• SAS thickening: 12.1%
• raw sludge dewatering: 5.2%
• anaerobic digestion and the subsequent digested sludge dewatering: 5.2%
As can be seen in Figure 5-8, the observed results match the model predictions well.
105
Figure 5-8 Comparison between the observed results and predicted results of the modelling of sludge treatment
(A) Primary sludge thickening, (B)Raw sludge dewatering, (C) SAS thickening, and (D) Anaerobic digestion
and digested sludge dewatering
Hence, the models constructed for modelling the sludge treatment including thickening,
dewatering and anaerobic digestion are considered representative of the actual performance
and the material mass balance of the WWTPs. The predicted flowrate and concentrations are
therefore used in the mass and energy balance calculations of the sludge treatment in this
chapter and Chapter 6.
The observed and predicted result of the sludge treatment modelling are shown in Table E-2 in
Appendix E.
106
5.2 The COD mass balance and chemical energy balance of treatment processes
5.2.1 The COD mass balance and chemical energy balance of primary settlement units
The flow and concentration data predicted via modelling were used in the construction of the
COD mass balance and chemical energy balances of primary settlement units.
In the four study WWTPs, the chemical energy of both wastewater and sludge have strong
linear relationships with COD (stated in Section 4.1.3 and 4.2.3). Therefore, the chemical
energy of the wastewater and sludge are estimated from the predicted COD (from modelling)
via Equation 4-2 (shown in Section 4.1.3).
In the case of WWTP A and D, which have reactive settlers, the outflowing COD mass and
chemical energy within the sludge and effluent wastewater is >1% smaller than the inflowing
COD mass and chemical energy. The missing COD and chemical energy are considered as loss
and are considered being caused by the biological reactions. However, the loss is less than 1%
Therefore even though biological sludge is recirculated back to the primary settlement tanks,
the COD loss caused by the biological reactions can be ignored. In the case of WWTPs B and
C, which have no biological sludge is recirculated back to the primary settlement tanks, the
outflowing COD mass and chemical energy within the sludge and effluent wastewater equals
the inflowing.
The COD mass and the chemical energy flows through the primary settlement units of all four
treatment plants can be balanced Hence, the COD mass and the chemical energy balances are
considered reliable. The COD mass balance and the chemical energy balance of the primary
settlement units of all four WWTPs are shown in Figure 5-9. In Figure 5-9, the C.E. presented
stands for chemical energy.
107
Figure 5-9 The COD mass and chemical energy balances of primary settlement in a daily basis at the four
WWTPs investigated
Typically COD removal efficiencies for primary settlement tanks are approximately 30-40 %,
with the remaining 60-70% of COD flowing to secondary treatment units (Wan et al., 2016,
Shi, 2011, Yeshi et al., 2013). The COD removal efficiencies of WWTPs B, C and D are similar
to this typical value. However, WWTP A removes about 60% of the COD in the primary
settlement process. This is probably because WWTP A doses ferric sulfate to its screened raw
wastewater. Although the ferric dosing solely aims to remove the phosphate (personal
communication, J McCowan, Scottish Water, April 2016), it may improve the sedimentation of
colloidal or particular matter and therefore potentially settles a greater proportion of COD from
the raw wastewater into the primary sludge (Ismail et al., 2012, Tchobanoglous et al., 2003).
As shown in Table 5-3, a higher pCOD removal leads to a higher COD removal. However,
there do not appear to be any relationships between the hydraulic retention time (HRT), the
ratio of pCOD:COD, the COD, sCOD or pCOD concentration and the efficiency of COD
removal in the primary tanks (shown in Figures 5-10 and 5-11), though this is based on rather
limited data and would therefore need further investigation.
108
Table 5-3 The predicted influent and effluent CODs concertation and the CODs removal efficiency of primary
settlement in four WWTPs
pCOD in raw wastewater, mg/L 786.8 478.4 467.7 218.9
pCOD in primary effluent, mg/L 206.4 197.8 276.9 92.9
pCOD removal efficiency* 73.8% 58.7% 40.8% 57.5%
sCOD in raw wastewater, mg/L 274.6 248.7 197.3 117.7
sCOD in primary effluent, mg/L 200.1 187.4 180.7 87.1
sCOD removal efficiency* 27.1% 24.6% 8.4% 26.0%
COD removal efficiency* 61.7% 47.0% 31.2% 46.5%
*: Removal efficiency = (Influent Concentration – Effluent Concentration) ÷ Influent Concentration
1200 80.0%
70.0%
1000
60.0%
pCOD:COD, %
CODs, mg/L
800
50.0%
600 40.0%
30.0%
400
20.0%
200
10.0%
0 0.0%
25.0% 35.0% 45.0% 55.0% 65.0%
Percentage COD captured in primary sludge
COD, mg/L sCOD, mg/L pCOD, mg/L pCOD:COD, %
Figure 5-10 The impact of CODs concentration and fractions to the percentage COD captured in primary sludge
12
10
8
HRT, hour
0
25.0% 35.0% 45.0% 55.0% 65.0%
Percentage COD captured in primary sludge
Figure 5-11 The impact of HRT to the percentage COD captured in primary sludge
109
In summary, for the four WWTPs investigated in this study, primary settlement captured 46.3
± 12.1% of influent COD and chemical energy in the primary sludge, with the remainder
flowing to the secondary treatment processes.
5.2.2 The COD mass balance and chemical energy balance of secondary biological treatment
(activated sludge)
COD mass balance and chemical energy balances of secondary biological treatment as
activated sludge. The chemical energy of the wastewater and sludge are estimated from the
predicted COD (from modelling) via Equation 4-2 (shown in Section 4.1.3).
In the four study WWTPs, the COD mass and chemical energy within the effluent sludge and
wastewater flows into activated sludge process is substantially smaller than the inflowing COD
mass and chemical energy. Considering the activated sludge can remove organic matter via
oxidation (Tchobanoglous et al., 2003), the missing COD mass and chemical energy is due to
the organic carbon being oxidized to CO2 as gaseous product and emitted to the atmosphere.
In the case of WWTPs B and C, which do not have biological reactive final clarification, the
outflowing and inflowing COD mass and chemical energy of the final clarification balances.
But in the case of WWTPs A and D, which have biological reactive final clarification, the
outflowing COD mass and chemical energy within the effluent sludge and wastewater flows
into final clarification is 1-2% smaller than the inflowing. The possible reason for this is that
some of the COD may be consumed by denitrification occurring in the final clarifiers (Yeshi et
al., 2013, Tchobanoglous et al., 2003). Considering the loss of COD mass and chemical energy
in the final clarifier is small, the activated sludge process and the final clarification process are
combined for the calculation of mass and energy balance. The COD mass balance and chemical
energy balances for the four WWTPs are shown in Figure 5-12. In Figure 5-12, the C.E.
presented stands for chemical energy.
110
Figure 5-12 The COD mass and chemical energy balances of activated sludge in a daily basis at the four
WWTPs investigated
Nowak et al. (1999) suggested that 25.0 - 60.0 % of the COD removed was stored in new
growth biomass, and the rest was lost as CO2. Hence, the ratio of COD in surplus activated
sludge (SAS, displays as sludge in Figure 5-12) and COD lost as CO2 was in the range 0.33
(25%:75%) to 1.50 (60%:40%). From the data shown in Figure 5-12, the calculated ratios of
WWTPs A (1.30), B (0.34), C (1.35) and D (1.32) fall in this range. One possible reason for
this high ratio in WWTP B was that both biomass assimilation and bio-flocculation participated
in capturing the COD into the SAS (Jimenez et al., 2015). The bio-flocculation captures the
COD without oxidizing it and can be enhanced via iron dosing (Jimenez et al., 2015). The raw
111
wastewater of WWTP B possibly has a high iron content because the part of the wastewater
could be from the runoff from abandoned metal mine (personal communication, L Wilkinson,
Northumbrian Water, 29 October 2018). Hence, it is possible that unexpected bio-flocculation
took place in WWTP B.
In summary, this study finds that activated sludge captures 44.4 ± 15.5% of influent COD and
chemical energy in the SAS. Simultaneously, 42.2 ± 13.0% of the influent COD and chemical
energy is lost, potentially as CO2. The remaining 13.7± 5.4% will leave the treatment works as
final effluent if there is no tertiary treatment.
5.2.3 The COD mass balance and chemical energy balance of sludge thickening and
dewatering facilities
COD mass balance and chemical energy balances of sludge thickening and dewatering facilities.
The TSS, VSS, COD mass balance and chemical energy balance are constructed for all four
sludge thickening and dewatering processes of WWTP C. All the mass flows are balanced, as
shown in Figure 5-13. In Figure 5-13, the C.E. presented stands for chemical energy.
Figure 5-13 The TSS, VSS, COD mass balances and chemical energy balance of the sludge thickening and
dewatering processes in WWTP C
112
TS and VS were measured in the sampling campaign, but not the TSS and VSS. Here, the
predicted TSS and VSS are used as reference values to demonstrate the distribution of the
particulate material.
In GPS-X, VSS contributes to pCOD (Hydromantis, 2017). As shown in Figure 5-4, in the
thickening of primary sludge and SAS, the distribution pattern of the VSS and COD are in
good agreement. But in the raw sludge and digested sludge dewatering, the percentage of the
total COD in the centrate of the raw centrifuge and the digested centrifuge is substantially
greater than the percentage of total VSS. That is likely due to the large proportion of COD in
centrate present as sCOD (as shown in Table 5-4). The sCOD is produced from the
solubilization of the particulate materials due to the hydrolysis and fermentation after long
storage times or digestion prior to the sludge dewatering (Tchobanoglous et al., 2003, Ubay-
Cokgor et al., 2005). In the case of primary sludge thickening and raw sludge dewatering,
although similar percentages of the VSS ends in the filtrate or centrate, because the raw sludge
contains more sCOD, a greater percentage of the COD ends in the raw sludge centrate than in
the primary sludge filtrate.
Table 5-4 The predicted sCOD and COD concentration in different filtrates and centrates
sCOD, mg/L COD, mg/L sCOD:COD
Primary Sludge Filtrate 2661.9 9777.2 27.2%
SAS Filtrate 47.8 1603.0 3.0%
Raw Sludge Centrate 6027.2 10536.7 57.2%
Digested Sludge Centrate 4059.1 6722.3 60.4%
The thickening of primary sludge, SAS and the dewatering of raw sludge take place before
digestion. Although approximately 80% or more of COD and chemical energy is captured in
the thickened and dewatered sludge, since part of the COD and chemical energy ends in the
filtrates or centrates, it suggests that not all the COD and chemical energy captured in the
primary sludge and secondary sludge will reach the digester. Therefore, sludge thickening and
dewatering appears to reduce the chemical energy input to the digester.
113
5.2.4 The COD mass balance and chemical energy balance of anaerobic digestion
COD mass balance and chemical energy balances of anaerobic digestion.
The percentage difference between the inflowing and outflowing COD and chemical energy
balances is no greater than ± 7.0% (Figure 5-14). Therefore, both the COD mass balance and
chemical energy balances appear to be reliable. However, unlike in the as the primary
settlement (Section 5.2.1), activated sludge (Section 5.2.2) and sludge thickening and
dewatering (Section 5.2.3), the chemical energy balance no longer fully follows the COD
balance as the percentage of chemical energy in the biogas is less than the percentage of COD.
Gas
COD:62.6%
(62.1 tonnes)
C.E.:56.2%
(244.5 MWh)
Digester Digested
Feed
Anaerobic Sludge
COD:100% COD:37.4%
Digestion (37.0 tonnes)
(99.1 tonnes)
C.E.:100% C.E.:37.4%
(435.0 MWh) (162.5 MWh)
(C.E. stands for chemical energy.)
Figure 5-14 The COD and chemical energy balance of anaerobic digestion in WWTP C
The chemical energy balance suggests that 56.2 % of the chemical energy in the sludge is
recovered as CH4. Mills et al. (2014) suggested that a typical advanced AD (AAD) process is
able to recover 51.0 % of the sludge energy into CH4, and Barber (2016) reported a figure of
60.0 %. The findings of this study therefore approximately match these literature values.
In this study, the outflowing chemical energy within the bio-methane and in the digested sludge
is 6.4% less than the inflowing. Mills et al. (2014) suggested a similar pattern: the outflowing
chemical energy within the bio-methane and in the digested sludge is 5.2% less than the
inflowing. In theory, the chemical energy loss during the anaerobic digestion is possible. This
thesis uses glucose (C6H12O6) as an example (shown in Example 5-1) to demonstrate this
phenomenon in theory.
114
Example 5-1
C6H12O6 = 3CH4 + 3CO2
C6H12O6(s) + 6O2(g) = 6CO2(g) + 6H2O(lq)
-1281 0 -393.51×6 -285.83×6 △rH=-2,804.04 kJ/mol
3CH4(g) + 6O2(g) = 3CO2(g) + 6H2O(lq)
-74.81×3 0 -393.51×3 -285.83×6 △rH=-2,671.08 kJ/mol
As shown in Example 5-1

• The complete anaerobic digestion of 1 mole of C6H12O6 produces 3 moles of CH4 and
3 moles of CO2. A complete oxidation of 1 mole of C6H12O6 and 3 moles of CH4
consumes the same amount of oxygen (O2).
• Based on the enthalpy calculation, although 3 moles of CH4 has the same amount of
COD as 1 mole of C6H12O6, the heat release from the combustion (△rH) of 3 moles of
CH4 is only 2,671.08 ÷ 2,804,04 = 95.2% of the heat released from the combustion of
1 mole of C6H12O6.
The result found in study (shown in Figure 5-14) and reported by Mills et al. (2014) agrees
with the theory.
115
5.3 Energy recovery opportunities based on WWTP mass balances
5.3.1 Energy product of wastewater treatment
Based on the results presented in Sections 5.2.1 and 5.2.2, the COD and chemical energy mass
balances are shown in Figure 5-15.
Figure 5-15 The COD and chemical energy mass balance of the four WWTPs investigated
WWTPs can be considered as reactors that convert their feedstock, the energetic raw
wastewater, into three outputs: (1) sludge, (2) treated wastewater as final effluent, and (3)
gaseous emissions. From an energy recovery perspective, not every output can be considered
as a product. What is critical is whether the output contains considerable amounts of chemical
energy and whether there is potential to recover it.
116
As shown in Figure 5-15, apart from WWTP C no more than 15.0% the COD and chemical
energy flowing into the WWTPs is present in the final effluent. Meanwhile, the COD
concentration of the final effluents are lower than 100 mg/L, which is too low for energy self-
sufficient recovery, i.e. the energy consumption for energy recovery would be greater than the
value of the energy recovered(Stoll et al., 2018). The final effluent is not therefore considered
to be a potential product at any of the WWTPs. The gas produced from the energetic organic
carbon is mainly CO2 which contains no chemical energy, and hence it is not a product neither
(Tchobanoglous et al., 2003). In these four WWTPs, 31.2 ± 14.5% of COD and chemical
energy flowing into the works is lost via a combination of oxidation in the aerobic activated
sludge process and final effluent discharge.
In current practice energy recovery from wastewater is commonly achieved via anaerobic
digestion of sludge (Pearce et al., 2014), in which case sludge is the only energy product.
However, the energy cannot be produced directly from the sludge. Rather, it is an intermediate
product, in the same way that gasoline can be combusted for producing heat or to drive a vehicle.
In the four study WWTPs, 68.6 ± 14.4% of COD and chemical energy flowing into the works
is captured in the sludge and sent for digestion.
5.3.2 Energy product of sludge treatment
In the four WWTPs investigated only WWTP C has anaerobic digestion onsite for energy
recovery from sludge. Therefore, the sludge treatment of WWTP C is used here for illustrative
purposes. Sludge treatment can be considered as a reactor. Its feedstock is the sludge, and the
outputs are the biogas, filtrates, centrates and digested cake. The COD mass and chemical
energy balance are constructed based on the predicted result obtained via modelling and are as
shown in Figure 5-16. As can be seen, the outflowing COD and chemical energy to the sludge
is smaller than the inflowing. The loss of COD and chemical energy during the sludge treatment
is likely due to the biological reactions occurred in the sludge storage and digestion.
117
Figure 5-16 The COD and chemical energy balance of the sludge treatment in WWTP C
(The value shows in the () are the percentage mass balance using the total COD and chemical energy inputted to
the sludge treatment as reference)
As shown in Figure 5-16, WWTP C treats the indigenous sludge and also the imported wet
sludge and imported dewatered sludge. The imported sludge contains the equivalent of 17.0.0%
of the inflowing COD and chemical energy to the works. Meanwhile, the imported dewatered
sludge contains the equivalent of 6.3% of the inflowing COD and chemical energy to the works.
Unlike the wastewater treatment, the sludge treatment distributes its influent COD and
chemical energy to its three outputs more evenly:
• 21.3 % of the sludge COD and sludge chemical energy ends in the dewatered digested
sludge
• 31.1% of the sludge COD and sludge chemical energy is retained in the filtrates and
centrates
• 43.7% of the sludge COD and 38.4% of the sludge chemical energy is recovered as CH4.
Although dewatered digested sludge, filtrates and centrates contain considerable amounts of
chemical energy, they are not available for direct energy use. Currently, at WWTP C, the
filtrates and centrates are recirculated back to the head of the works, and the dewatered digested
cake is sent for agricultural land use (personal communication, L Wilkinson, Northumbrian
118
Water, October 2015). Therefore, CH4 is the only energy product of the combined wastewater
and sludge treatment. It is worth noting that the percentage sludge COD retained in the filtrates
and centrates are substantially greater that the value reported by Shi (2011) (17.1%) and Wett
et al. (2007) (2.4%).
As shown in the Figure 5-16, the COD and energy outputs do not equal the inputs. That is
because part of the COD and chemical energy is lost during sludge storage and digestion but is
not accounted for in Figure 5-16.
5.3.3 Energy recovery potential from wastewater
5.3.3.1 Chemical energy recovery potential of wastewater

If the wastewater treatment and sludge treatment are considered as two reactors in sequence,
the chemical energy that could potentially be generated (ECEGen) from 1 m3 of wastewater can
be empirically estimated by Equation 5-1.
Equation 5-1
𝐸𝐶𝐸𝐺𝑒𝑛 = 𝐸𝑤𝑤 × 𝜂𝑆𝑓𝑊𝑊 × 𝜂𝐶𝐸𝑓𝑆
Where 𝑬𝒘𝒘 is the chemical energy content of 1 m3 of wastewater and can be estimated via
multiplying its COD concentration by 0.0158 kJ/mg COD (0.0044 kWh/g COD),
𝜼𝑺𝒇𝑾𝑾 is the percentage of the chemical energy flowing into the works that is captured
in the indigenous sludge, and
𝜼𝑪𝑬𝒇𝑺 is the percentage of the sludge chemical energy recovered in the energy product
from the sludge treatment.
𝜂𝐶𝐸𝑓𝑆 is affected by the two factors;

1. the percentage of sludge chemical energy that could reach the digester after several
stages of thickening and dewatering (𝜂𝐶𝐸𝑟𝐷 )
2. the energy recovery efficiency of the digester ( 𝜂𝐶𝐸𝑓𝐷 ). 𝜂𝐶𝐸𝑓𝑆 can be expressed as
Equation 5-2.
Equation 5-2
𝜂𝐶𝐸𝑓𝑆 = 𝜂𝐶𝐸𝑟𝐷 × 𝜂𝐶𝐸𝑓𝐷
119
Because WWTPs A, B and D has no sludge digestion, and hence these WWTPs has no studied
𝜂𝐶𝐸𝑟𝐷 and 𝜂𝐶𝐸𝑓𝐷 . In the estimation of the ECEGen from the raw wastewater of WWTPs A, B and
D, the universal values of 𝜂𝐶𝐸𝑟𝐷 and 𝜂𝐶𝐸𝑓𝐷 are used based on the chemical energy balance of
the sludge treatment WWTP C. As shown in Figure 5-16, 91.9 units of sludge energy are first
fed to the sludge treatment, but after several stages of thickening and dewatering only 60.2
units reaches the AAD. The 𝜂𝐶𝐸𝑟𝐷 is 65.5%. After 8.1 units of sludge energy from the imported
dewatered sludge is added, the AAD receives 68.3 units of sludge energy and generates CH4,
which is equivalent to 38.4 units of sludge energy. Thus, the 𝜂𝐶𝐸𝑓𝐷 is 56.2%. The ECEGen values
of all four WWTPs are estimated from the COD concentration of the raw wastewater (shown
in Table 5-1) via Equation 5-1 and listed in Table 5-5.
Table 5-5 The estimated chemical energy production potential from the four study WWTPs
COD, mg/L Eww, kWh/m3 𝜼𝑺𝒇𝑾𝑾 ** ECEGen, kWh/m3 ECEGen:Eww
WWTP A 1061.4 4.66 76.8% 1.32 28.3%
WWTP B 727.1 3.19 82.9% 0.97 30.5%
WWTP C 665.0 2.92 54.9% 0.59 20.2%
WWTP D 336.6 1.48 65.0% 0.35 23.9%
*: The percentage chemical energy capture in the sludge shown in Figure 5-15
As shown in Table 5-5, both the COD of the raw wastewater and the 𝜂𝑆𝑓𝑊𝑊 influence the
ECH4Gen. In the case of WWTP C and WWTP D, although WWTP D has a higher 𝜂𝑆𝑓𝑊𝑊 than
WWTP C (65.0% compared to 54.9%), since the COD of WWTP C raw wastewater is
approximately double that of WWTP D, the ECEGen of WWTP C is higher than ECH4Gen of
WWTP D. In the case of WWTP B and C, although the COD of their raw wastewater is similar,
since the 𝜂𝑆𝑓𝑊𝑊 of WWTP B is about 1.5 times greater than WWTP C, the ECEGen of WWTP
B is much higher than that of WWTP C (0.97 kWh/m3 compared to 0.59 kWh/m3). The values
of ECEGen:Eww in Table 5-4 suggest that only approximately 20-30% of the chemical energy
contained in the raw wastewaters of these WWTPs can be recovered as CH4 energy. In reality,
the ECEGen:Eww of WWTP C is 30.0% (shown in Figure 5-16). This higher ratio is attributed to
the extra sludge imported to the sludge treatment.
5.3.3.1 Electrical energy recovery potential of wastewater

Wastewater and sludge treatment have the potential to produce energy, but they also consume
energy. Energy used in wastewater treatment is usually in the form of electrical energy, and for
120
sludge treatment both electrical and heat energy may be used (Wan et al., 2016, Mills et al.,
2014). However, the energy balance of a WWTP is usually expressed as a function of kWh
electricity used per m3 of wastewater. Current practice in the wastewater treatment sector is to
recover electrical energy from wastewater by feeding the CH4 produced in sludge digestion to
an electricity generator (Rawlinson et al., 2012, Gu et al., 2017). To construct a more
meaningful energy balance it is therefore useful to convert the energy of wastewater and sludge,
in units of heat or chemical energy, into units of electrical energy, so that a direct comparison
of electricity consumed to potential electricity available can be made for individual wastewater
treatment plants (Wan et al., 2016). For this research this has been done by combining the total
chemical energy potentially available in the wastewater and sludge treatment processes, thus
producing a single figure for available electricity.
One of the key tasks to implement this approach is to convert the heat energy consumed during
the advanced anaerobic digestion process into an expression of electrical energy, because the
THP process requires considerable amounts of heat input. In practice the heat is commonly
supplied via two approaches. As shown in the Figure 5-17, the heat is supplied from the heat-
grade heat recovered from the Combined Heat and Power (CHP) generation. If insufficient heat
is provided by the CHP the deficit is supplied from a boiler that uses supportive fuel, such as
natural gas or oil (Bowen et al., 2010). Thus, the input to the system is not only the sludge
energy, but also the chemical energy from the support fuel. This energy usage needs converting
to electrical energy in the approach used for the energy mass balance in this research. The
second approach to the provision of heat, is to first divert some of the CH4 to the boiler, and
then the heat demand is met by combining the heat production from both the boiler and the
CHP (Smyth et al., 2016). In this approach no support fuel is needed. There is only one input
to the system in this case. The ratio for splitting the CH4 to the boiler and the CHP has to be
calculated because it determines the quantity of CH4 available to the CHP.
121
(A)
Input
Boiler Supportive
Fuel
Combined
Thermal Anaerobic Output
Heat &
Hydrolysis Digestion Electricity
Power
(B)
Boiler
Combined
Input Thermal Anaerobic Output
Heat &
Sludge Hydrolysis Digestion Electricity
Power
Sludge Biomethane/Natural High-grade

gas heat
Figure 5-17 Schematic illustration of different approaches to heat supply to the THP
The ratio for splitting the CH4 to the boiler and CHP for achieving heat self-sufficient can be
calculated, beginning with the heat demand of the THP (𝜂𝐻𝑡𝑇𝐻𝑃 ). The heat demand of THP
(𝜂𝐻𝑡𝑇𝐻𝑃 ) is considered as the unit heat required per unit sludge energy fed to the THP treatment
unit and the subsequent digestion. Since the demand is met by high-grade heat from the CHP
and the heat from the boiler, 𝜂𝐻𝑡𝑇𝐻𝑃 can be calculated using Equation 5-3.
122
Equation 5-3
𝜂𝐻𝑡𝑇𝐻𝑃 = 𝜂𝐶𝐸𝑓𝐷 × 𝜂𝐶𝐸𝑡𝐶𝐻𝑃 × 𝜂𝐻𝐺𝐻𝑓𝐶𝐻𝑃 + 𝜂𝐶𝐸𝑓𝐷 × (1 − 𝜂𝐶𝐸𝑡𝐶𝐻𝑃 ) × 𝜂𝐵𝑜𝑖𝑙𝑒𝑟
Where 𝜼𝑪𝑬𝒕𝑪𝑯𝑷 is the proportion of the produced chemical energy fed to the CHP
𝜼𝑪𝑬𝒇𝑫is the chemical energy recovery efficiency of the digester,
𝜼𝑯𝑮𝑯𝒇𝑪𝑯𝑷 is the high-grade heat recovery efficiency from per unit energy fed
to the CHP,
𝜼𝑩𝒐𝒊𝒍𝒆𝒓 is the thermal efficiency of the boiler.
The 𝜂𝐶𝐸𝑡𝐶𝐻𝑃 can be calculated using Equation 5-3 which is rearranged from Equation 5-4.
Equation 5-4
𝜂𝐻𝑡𝑇𝐻𝑃 − 𝜂𝐶𝐸𝑓𝐷 × 𝜂𝐵𝑜𝑖𝑙𝑒𝑟
𝜂𝐶𝐸𝑡𝐶𝐻𝑃 =
𝜂𝐶𝐸𝑓𝐷 × (𝜂𝐻𝐺𝐻𝑓𝐶𝐻𝑃 − 𝜂𝐵𝑜𝑖𝑙𝑒𝑟 )
Historical data shows 77.3 MWh/d heat as steam is used for the THP to treat feed sludge to
anaerobic digestion process during the period of the sampling campaign conducted (personal
communication, S Coverdale, Northumbrian Water, October 2017). As shown in Figure 5-14,
the digester feed sludge contains 435.0 MWh chemical energy. Thus, the 𝜂𝐻𝑡𝑇𝐻𝑃 (heat demand
of THP) of the sludge treatment of WWTP C is 77.3÷435.0 = 17.8% which is in agreement
with the reported range of 13.6-17.2% (Smyth et al., 2016, Mills et al., 2014). A value for
𝜂𝐻𝐺𝐻𝑓𝐶𝐻𝑃 (heat recovery efficiency) of 20.0% is assumed since the reported range is 18-22%
(Mills et al., 2014, Smyth et al., 2016, Bowen et al., 2010). The boiler efficiency (𝜂𝐵𝑜𝑖𝑙𝑒𝑟 ) is
taken the reported value of 85% (Vanwortswinkel and Nijs, 2010). From this, the 𝜂𝐶𝐸𝑡𝐶𝐻𝑃 is
calculated as 0.821. To achieve self-sustaining heat generation at WWTP C, theoretical
calculations suggest that 82.1% of the chemical energy produced via THP (as CH4) at the plant
should be fed to the CHP for electricity generation.
With a common heat to electricity conversion efficiency (𝜂𝐸𝑓𝐶𝐻𝑃 ) of 38% in Combined Heat
and Power (CHP) electricity generation (Banks, 2009, Mills et al., 2011), the electricity that
could be produced from 1 m3 of wastewater (EElecGen) can be calculated using Equation 5-5.
123
Equation 5-5
𝐸𝐸𝑙𝑒𝑐𝐺𝑒𝑛 = (𝐸𝑤𝑤 × 𝜂𝑆𝑓𝑊𝑊 × 𝜂𝐶𝐸𝑟𝐷 × 𝜂𝐶𝐸𝑓𝐷 ) × 𝜂𝐶𝐸𝑡𝐶𝐻𝑃 × 𝜂𝐸𝑓𝐶𝐻𝑃
The estimated EElecGen of WWTP A, C and D, and their electrical balances are shown in Table
5-6.
Table 5-6 The estimated electricity production of different wastewater
WWTP electricity consumption,
Flow, m3/d* EElecCon, kWh/m3 EElecGen, kWh/m3
kWh/d**
WWTP A 9,052 3,015 0.33 0.41
WWTP C 278,954 85,949 0.31 0.18
WWTP D 37,312 9,744 0.26 0.11
*: It is shown in Table 5-1
**: It is collected from Scottish Water and Northumbrian Water (personal communication, J McCowan, Scottish
Water, June 2017 and personal communication, L Wilkinson, Northumbrian Water, July 2018)
In case of WWTPs A, C and D, the chemical energy contained in 1 m3 raw wastewater (Eww,
1.48-4.66 kWh, Table 5-5) is approximately 5-14 times higher than the electricity consumed in
treating 1 m3 of wastewater (EElecCon, 0.26-0.33 kWh, Table 5-6). However, the estimated
EElecGen are 0.11-0.41 kWh (Table 5-6), only 6 - 9% of the Eww in the wastewater. With the
exception of WWTP A the potential EElecGen is less than EElecCon, and hence WWTPs C and D
do not have the potential to be electricity self-sustaining. Although the potential EElecGen of
WWTP A is greater than its EElecCon, considering WWTP A has no onsite anaerobic digestion,
if energy consumed on the sludge haulage is taken into account, the surplus EElecGen compared
to EElecCon will be reduced or even eliminated.
5.3.3.2 Improving the electrical energy balance

The wastewater and subsequent sludge treatment with anaerobic digestion is considered as a
set of reactors in sequence. The intermediate product is the sludge. The 𝜂𝑆𝑓𝑊𝑊 can be
considered as the efficiency with which COD or chemical energy is captured in the wastewater
treatment plant. A greater 𝜂𝑆𝑓𝑊𝑊 result in a greater amount of feedstock to the second reactor,
the sludge treatment with anaerobic digestion. The sludge produced by the wastewater
treatment consists of the primary sludge and SAS. However, the feed of the activated sludge is
the effluent of the primary settlement. If the COD capture rate of the primary settlement is X,
the COD capture rate of the activated sludge is Y, and the COD and chemical energy capturing
rate of the WWTP, 𝜂𝑆𝑓𝑊𝑊 , can be expressed by Equation 5-6.
124
Equation 5-6
𝜂𝑆𝑓𝑊𝑊 = 𝑋 + (1 − 𝑋)𝑌 = (1 − 𝑌)𝑋 + 𝑌
X, Y and 𝜂𝑆𝑓𝑊𝑊 are all between 0 to 1. According to the equation, if Y is given a set value, (1-
Y) becomes a constant that is bigger than 0, then the Equation 5-6 becomes a linear function
of X. Since the slope is bigger than 0, a greater X to can lead to greater 𝜂𝑆𝑓𝑊𝑊 . Improving the
primary settlement process can potentially increase the 𝜂𝑆𝑓𝑊𝑊 (Equation 5-5) and lead to a
greater EElecGen. Primary settlement aims to settle the particulate contaminant. In the four study
WWTPs, approximately 26-60 % of the particulate COD remained in effluent of primary
treatment (Table 5-3). The primary settlement can still be improved. Research suggests this can
be done by dosing aluminum and ferric salts at the correct dosage rate and pH (Ismail et al.,
2012, Sarparastzadeh et al., 2007).
The 𝜂𝐶𝐸𝑟𝐷 (Equation 5-5) is the energy recovery efficiency of the sludge treatment. This
consists of several stages of sludge thickening and dewatering prior to anaerobic digestion. As
stated in Section 5.2.3, filtrates and centrates contain 31.1% of sludge energy. Among all four
sludge filtrates and centrates, 1) primary sludge filtrate, 2) SAS filtrate, and 3) raw sludge
centrate are formed prior to the anaerobic digestion. The formation of the filtrate and centrate
does have an adverse effect on final energy production because sludge energy is kept in the
filtrate rather than being sent for recovery. Therefore, reducing the chemical energy contained
in the filtrates and centrates is a potential way of elevating the digestion feedstock and hence
improving CH4 production.
Recovering energy from the un-recovered chemical energy reserve is also a measure to promote
the EElecGen. From the chemical energy balance study of the sludge treatment (Figure 5-16), the
end product of the digestion, the dewatered digested cake, contains approximately 20% of the
influent sludge energy fed to the sludge treatment. Although it is currently not used for energy
application in WWTP C, has proven could produce energy via pyrolysis or gasification (Mills
et al., 2014, Cao and Pawłowski, 2012). The energy recovery from digested cake could also
leads to a greater energy recovery efficiency from sludge.
125
Apart from improving the energy balance of WWTPs by focusing on elevating the EElecGen,
efforts could also be made to reduce EElecCon. Activated sludge process, which is used by all
four WWTPs and is well applied in wastewater treatment, is reported to account for
approximately 50% of the total process energy demand (Scherson and Criddle, 2014). Although
aerobic activated sludge still produces SAS that can be used for energy production via digestion,
it consumes energy to convert the energy material into CO2 (Wan et al., 2016). As shown in
Figure 5-12, the activate sludge processes suffered 23-52% of chemical energy loss. Activated
sludge process is further wasting the chemical energy within the wastewater. Energy generating
biological treatment, such as anaerobic treatment and microbial fuel/electrolysis cell, generates
energy, for instance CH4, H2 and electricity, while simultaneously removing organic
contaminants (Shoener et al., 2014, Stoll et al., 2018). The produced energy is then used for
offsetting energy consumption. This could be an alternative to the current aerobic activated
sludge treatment.
5.4 Case studies of energy recovery
Four case studies are presented in this section. They aim to illustrate how it may be possible to
improve the electrical energy balance of the WWTP by:
1. Improving solids removal efficiency in the primary settlement (Section 5.4.1)

2. Reducing the sludge retention time prior to digestion (Section 5.4.2)
3. Recovering energy from the dewatered digested sludge via pyrolysis (Section 5.4.3), and
4. Replacing the activated sludge processes with an energy-generating anaerobic membrane
reactor (Section 5.4.4).
Apart from case study 3, all case studies are founded on software modeling outputs from GPS-
X. The conceptual models used are built on the model of WWTP C. That is because the model
has been calibrated and is able to represent the real-life performance of the sludge thickening,
dewatering and sludge digestion. As stated in Sections 5.3.3, the COD concentration of the raw
wastewater is a key influence on the prediction of the chemical energy and electricity that might
be recoverable from wastewater. Therefore, the case studies do not suggest the exact value of
chemical energy or electricity that could be recovered but illustrate the relative merits of these
approaches in terms of potential energy production and energy consumption.
126
It is worth noting that the electricity consumption and generation are not predicted via the GPS-
X. This is because historical data collected from Northumbrian Water for WWTP C does not
clearly show the electricity consumption of each individual unit process but shows electricity
usage for groups of processes (personal communication, L Wilkinson, Northumbrian Water,
July 2018). These include screening, odor control, wastewater pumping, primary settlement
scraping, RAS and SAS pumping, aeration, and sludge treatment (as shown in Figure 5-18).
Hence, it is not possible to model the electricity consumption of individual units in the software.
It is worth noting that the activated sludge the RAS and SAS pumping and aeration contributed
32% of the total consumption. Whilst, only 1% of the consumption is for the primary settlement
scrapping.
1758.0, 2%
5800.5, 6%
8684.0, 9%
7481.7, 7%
30553.8, 30%
12902.3, 13%
11782.6, 12%
20168.1, 20%
1302.8, 1%
Screening Wastewater Pumping

Primary settlement scrapping RAS and SAS Pumping
Aeration UV disinfection
Odour control Sludge treatment
Site administrative use
Figure 5-18 The breakdown of daily electricity consumption (kWh/d) of WWTP C
In this research, it is assumed that:

• Electricity consumption of primary settlement is proportional to the total solids content
of the primary sludge,
• Electricity consumption of aeration is assumed proportional to total air flow required,
• Electricity consumption for sludge treatment (including pumping, thickening and
dewatering) is proportional to the total TS fed to all four of the thickening and
dewatering facilities (Mills et al., 2014),
• The other usage is considered as fixed consumption that will not change.
127
The per unit electricity consumption of the primary settlement, aeration and sludge treatment
are calculated based on modelling results of the mass balance of WWTP C and are shown in
Table 5-7.
Table 5-7 The per unit electricity consumption of the primary settlement, aeration, sludge treatment and fixed
consumption of the WWTP C
Total Consumption,
Quantity Per Unit Consumption
kWh
Primary Settlement 1302.8 54.2 tonnes TS 24.0 kWh/ton TS
Aeration 20,168.1 2,316,000 m3 Air 0.0087 kWh/m3 Air
Sludge treatment 8,684.0 261.8 tonnes TS 33.2 kWh/ton TS
Fixed Consumption 70,279 N/A N/A
It is worth noting that the TS of the sludge is calculated from the predicted COD concentration
of the sludge using Equation 5-7 which is derived from Equation 4-4 (in Section 4.2.3).
Equation 5-7
VS
𝑇𝑆 = (𝐶𝑂𝐷 + 66.5) ÷ 1.3526 ÷ %
TS
In Table 4-6, the mean observed VS/TS% of primary sludge, SAS, raw sludge, and digested
sludge are 82.6±3.5%, 78.2±5.1%, 78.3±6.1%, and 67.8±1.6 %, respectively. For
simplification, the VS/TS% of the primary sludge, SAS and raw sludge is taken as 80% and
for digested sludge is taken as 70%.
5.4.1 Case study: Improving solids removal efficiency in primary settlement
5.4.1.1 Introduction
The sludge produced in the wastewater treatment becomes the feed stock to anaerobic digestion.
As stated in Section 5.3.3, if the kinetic biological reactions of the secondary treatment remain
constant, the wastewater treatment process will capture more chemical energy in the indigenous
sludge if the primary settlement process is able to settle more chemical energy in its sludge.
This study aimed to investigate the impact to the energy balance of the treatment works in a
scenario in which solid removal efficiency of the primary settlement is improved.
128
5.4.1.2 Method
A conceptual model CM-4-1 (Figure 5-19) was constructed based on the model of WWTP C.
Apart from transforming the model used for the primary settler from the “Simple1d” to an
“empiric” model, the flow and concentrations of the raw wastewater and the configurations of
the other processes remained the same. In the primary settler the settling model is changed to
“empiric” and the sludge discharged is controlled by both the solid removal efficiency and the
solid concentration of the primary sludge. The solid concentration is set to 26,360 mg/L which
is the same as that in WWTP C.
Figure 5-19 The conceptual model CM-4-1
For the sludge treatment modelling, it was assumed that there was no imported sludge, either
before primary sludge thickening or after the raw sludge centrifuge. For modelling the
digestion, to characterize the digester feed COD, TKN, TP, NH4+-N, PO43--P, VSS/TSS are
required input variables. The study fixed the COD concentration of the digester feed (CODDF)
at 113,13780,280 mg/L which is the same as the original model. The flow of the digester feed
(QDF) was calculated using Equation 5-8
Equation 5-8
𝑄𝐷𝑅𝑆 × 𝐶𝑂𝐷𝐷𝑅𝑆
𝑄𝐷𝐹 =
𝐶𝑂𝐷𝐷𝐹
Where QDRS is the predicted flow of the dewatered raw sludge
CODDRS is the predicted COD concentration of the dewatered raw sludge.
129
The TKN, TP, NH4+-N, PO43--P, VSS and TSS concentrations are then calculated using
Equation 5-9.
Equation 5-9
𝑄𝐷𝑅𝑆 × 𝐶𝐷𝑅𝑆
𝐶𝐷𝐹 =
𝑄𝐷𝐹
Where 𝑪𝑫𝑭 is the concentration of a parameter in the digester feed
𝑪𝑫𝑹𝑺 is the concentration of the corresponding parameter in the dewatered raw
sludge.
The solid removal efficiency of primary settlement (SREPS) was incrementally changed from
0.40 to 0.70 at an interval of 0.05, to directly evaluate the influence of improved primary
settlement efficiency on the energy balance. Changing the solid removal rate will affect the
volume of sludge produced, which in turn may change the HRT in the raw sludge tank and in
the digester. For the purposes of modelling, in order to make sure that the sludge undergoes
fermentation to approximately the same extent during hydrolysis and digestion, in this study
HRT was fixed at 410 hours and 505 hours in the raw sludge tank and the digester, respectively.
The kinetics and model stoichiometry of the biological reaction in the activated sludge, all the
sludge storage and anaerobic digestion will not be changed. Neither are the fractions of COD,
nitrogen and phosphorus of the raw wastewater and the digester feed. The operational
parameters at WWTP C were used in the models, and that they were always the same.
5.4.1.3 Results and discussion

The greater SREPS first results in a greater percentage of chemical energy captured in the
primary sludge. It also results in a smaller percentage of chemical energy flowing to the
activated sludge. In this case study, the stoichiometry and kinetics of the biological reactions
in the activated sludge process stays the same, a smaller percentage of chemical energy is
therefore captured in the SAS. However, the overall percentage chemical energy captured in
the combined sludge increases from 50.3% to 58.6%. More chemical energy from the
wastewater is captured and sent to the sludge treatment. The chemical energy balances are
shown in Table 5-8.
130
Table 5-8 The chemical energy balance of the models with different SREPS
SREPS
0.40 0.45 0.50 0.55 0.60 0.65 0.70
Final Effluent 15.1% 14.9% 14.9% 14.8% 14.8% 14.7% 14.7%
Loss in Activated sludge 34.7% 33.8% 32.5% 31.2% 29.7% 28.2% 26.7%
Primary Sludge 29.6% 33.2% 36.9% 40.6% 44.2% 47.9% 51.5%
SAS 20.7% 18.1% 15.7% 13.4% 11.3% 9.2% 7.1%
Primary sludge filtrate 2.8% 3.1% 3.4% 3.8% 4.1% 4.5% 4.8%
SAS filtrate 3.9% 3.1% 2.5% 2.1% 1.7% 1.3% 1.0%
Raw sludge centrate 8.1% 8.0% 7.7% 7.5% 7.2% 7.0% 6.6%
Dewatered Raw Sludge (as Digester Feed) 32.8% 34.7% 36.7% 38.7% 40.8% 42.9% 45.0%
CH4 21.3% 22.5% 23.8% 25.1% 26.5% 27.8% 29.2%
Dewatered Digested Sludge 9.7% 10.3% 10.9% 11.5% 12.1% 12.8% 13.4%
Digested Sludge Centrate 1.7% 1.8% 1.9% 2.0% 2.1% 2.2% 2.3%
As shown in Table 5-8, in the subsequent thickening, the greater SREPS increases the percentage
chemical energy retained in the primary sludge filtrate because the flow of primary sludge
increases and reduces the percentage chemical energy in the SAS filtrate because there is less
production of SAS. Because the reduction of chemical energy in the SAS filtrate is greater than
the increase in chemical energy in the primary sludge filtrate, when the SREPS increases, the
percentage chemical energy lost in the two combined filtrates still reduces from 6.7%
(SREPS=0.4) to 5.8% (SREPS=0.7).
In addition, with an increase in SREPS, the proportion of the proportion of SAS in the combined
sludge decreases. As shown in Table 5-9, when the SREPS increases, the predicted total COD
fed to raw sludge increases slightly. At the same time, the substrate COD can be increased up
to 45.7% whilst the heterotrophic biomass COD (originally produced in the activated sludge
process) can be reduced up to 52.2%.
131
Table 5-9 COD, Substrate COD and Heterotrophic biomass COD of the sludge fed to the raw sludge storage
tank
SREPS
0.40 0.45 0.50 0.55 0.60 0.65 0.70
COD of the sludge fed to the raw sludge
56280 57460 58350 59050 59670 60180 60570
storage tank, mg/L
Substrate COD of the sludge fed to the
27870 30050 32270 34450 36610 38660 40610
raw sludge storage tank, mg/L
Heterotrophic biomass COD of the sludge
13450 12630 11530 10310 9036 7746 6435
fed to the raw sludge storage tank, mg/L
The biological model used in the “Mantis2” model is developed from the activated sludge
model No.2 d (ASM2d) (Hydromantis, 2017). In ASM2d, the extent of hydrolysis and
fermentation are positively related to the concentration of heterotrophic biomass COD (Henze
et al., 1999). As the HRT of the sludge storage time is fixed, because less heterotrophic biomass
presents, hence less fermentation and hydrolysis happened in the raw sludge storage tank.
Therefore, less COD is converted into sCOD which would be lost in the subsequent sludge
dewatering. When the SREPS is changed from 0.4 to 0.7, the percentage COD loss due to raw
sludge dewatering is reduced from 8.1% to 6.6%, and the overall percentage COD loss prior to
digestion is 14.8% to 12.4%. Since more energetic COD is available to the digester, more
chemical energy is therefore available to the digester and leads to more biogas is produced.
Meanwhile, the extent of digestion of sludge is fixed in all cases, a greater digester feed leads
to a greater percentage chemical energy captured in the dewatered digested sludge.
With respect to electricity, although the increase of SREPS results in less energy consumption
during aeration (from 20,203 kWh/d to 16,180 kWh/d), the electricity consumption for primary
settlement and sludge treatment also increases due to a greater production of primary sludge
(from 6,417 kWh/d to 7,671 kWh/d). Overall, the electricity consumption is only reduced
slightly, by up to 1.9%. However, the electricity production could increase by up to 32.2%. The
electricity deficit is substantially reduced, and the reduction is up to 39.3%. The electricity
balance is shown in Table 5-10.
132
Table 5-10 The electricity balance of the models with different SREPS
SREPS
0.40 0.45 0.50 0.55 0.60 0.65 0.70
Primary Settlement, kWh/d 1217 1368 1519 1670 1821 1971 2122
Aeration, kWh/d 20203 20046 19541 18836 18026 17129 16180
Fixed consumption, kWh/d 70279 70279 70279 70279 70279 70279 70279
Sludge Treatment, kWh/d 6417 6572 6763 6974 7196 7433 7671
Total Consumption, kWh/d 98116 98265 98102 97759 97322 96812 96252
Electricity Generation, kWh/d 51352 53619 56348 59251 61900 64944 67872
Electricity Deficit, kWh/d 46764 44647 41754 38507 35422 31868 28380
Improving the SREPS may result in elevated biogas production, slightly reduced electricity
consumption, and consequently an improvement in the overall energy balance of the
wastewater treatment works.
The uncertainty in this case study is how to actually improve the SREPS operationally. It has
been suggested that SRE can be improved by dosing aluminum and ferric salts (Ismail et al.,
2012, Sarparastzadeh et al., 2007). However, in this case study the SREPS is adjusted by tuning
the operational parameter of the “empiric” model that controls the primary settlement. There is
no chemical dosing involved since the chemical dosing in the GPS-X is not able to enhance the
SREPS. Meanwhile, Diamantis et al. (2013) reports that a high dosage of aluminium or ferric
coagulant may reduce the digestibility. Therefore, if chemical dosing is applied to elevate the
SREPS, a thorough investigation of possible changes to the dewaterability or digestibility of the
sludge is recommended.
5.4.2 Case study: Reducing the sludge retention time prior to digestion
Sludge thickening and dewatering aims to reduce water content of the feed sludge for various
purposes, such as reducing sludge volume or raising the TS content to meet requirements for
subsequent treatment. Inevitably some of the solid, both volatile and fixed, will remain in the
filtrate or centrate rather than being retained in the thickened or dewatered sludge. From a
waste recovery perspective, considering that the sludge will ultimately be fed to the digester
for energy recovery, that portion of the solid that remains in the filtrate or centrate will not be
133
available for energy production. The loss of solids in the filtrate or centrate should therefore be
minimized.
The loss of solid can be influenced by the solids retention capability of the sludge thickening
and dewatering process, but also by biological activity in the sludge before thickening and
dewatering. In the case of WWTP C, considerable amounts of the particulate COD of the feed
sludge to the raw centrifuge are solubilized due to the long storage time, effectively resulting
in greater solid loss in the centrate. This case study investigates the impact of reducing the size
of the sludge storage tank to reduce the loss of potential energy during the raw sludge
dewatering.
5.4.2.2 Method
The conceptual model CM-4-1 is applied in this case study. The sludge retention time of the
raw sludge storage tank was decrementally changed from its original level of 410 hour to 0
hour at a percentage interval of 20%, to directly evaluate the influence of sludge storage time
on the energy balance. The raw sludge storage tank was therefore assumed as a virtual unit. Its
tank size was given values of 26,086 m3 (100%), 20,869 m3(80%), 15,652 m3(60%), 10,434
m3(40%), 5,217 m3(20%), 0 m3(0%). When the tank size is 0 m3, it means the raw sludge will
not be stored.
As in Section 5.4.1, the tank size of the digester will be adjusted to ensure approximately
constant HRT. The solid removal efficiency of the primary settlement tank will also be fixed at
0.40. The kinetics and model stoichiometry of the biological reaction in the activated sludge,
all the sludge storage tanks and anaerobic digestion will not be changed. The fractions of COD,
nitrogen and phosphorus of the raw wastewater and the digester feed also remain the same.
Finally, the operational parameters of the sludge dewatering unit remain unchanged in this
study.

The chemical energy balance of the wastewater treatment, primary sludge thickening, and SAS
thickening is not influenced by the change in volume of the raw sludge storage tank.
134
When the effects of a decrease in volume of the raw sludge storage tank are modelled, the
retention time of sludge reduces, and hence the degree to which raw sludge solubilizes
decreases. As a result, the sCOD of the raw sludge centrate decreases. As shown in Table 5-11,
there may be an approximately 62-fold variation in sCOD concentration, with a consequent
variation of the COD in the raw sludge centrate.
Table 5-11 The change of sCOD, pCOD and COD over different raw sludge storage time
Tank Volume, m3 sCOD, mg/L pCOD, mg/L COD, mg/L
26086 7305 5073 12378
20869 6628 5247 11875
15652 6351 5422 11773
10434 5631 5695 11326
5217 3719 6128 9847
0 117 6872 6989
The reduction of tank size reduces the percentage COD (also chemical energy) lost in the raw
sludge centrate and simultaneously increases the percentage of COD (and chemical energy) fed
to the digester, from 34.9% up to 41.4%. Ultimately, this leads to a greater biomethane
production up to 23.3% of the chemical energy can be recovered as. The chemical energy
balances are shown in Table 5-12.
Table 5-12 The chemical energy balance of models with different raw sludge storage time
Tank volume, m3
26086 20869 15652 10434 5217 0
Final Effluent 13.7% 13.7% 13.7% 13.7% 13.7% 13.7%
Loss in Activated sludge 31.6% 31.6% 31.6% 31.6% 31.6% 31.6%
Primary Sludge 29.5% 29.5% 29.5% 29.5% 29.5% 29.5%
SAS 25.3% 25.3% 25.3% 25.3% 25.3% 25.3%
Primary Sludge Filtrate 2.8% 2.8% 2.8% 2.8% 2.8% 2.8%

SAS Filtrate 5.7% 5.7% 5.7% 5.7% 5.7% 5.7%
Raw Sludge Centrate 8.7% 8.3% 8.2% 7.9% 6.9% 4.9%
Dewatered Raw Sludge (as Digester Feed) 34.9% 35.5% 36.3% 37.4% 39.0% 41.4%
CH4 19.6% 19.9% 20.4% 21.0% 21.9% 23.3%
Dewatered Digested Sludge 10.7% 10.9% 11.2% 11.5% 12.0% 12.8%
Digested Sludge Centrate 1.7% 1.7% 1.8% 1.8% 1.9% 2.0%
With respect to electricity, a decrease in size of the raw sludge storage tank has no impact on
the processes prior to the raw sludge dewatering, as shown in Table 5-13. Because solids
135
solubilization is reduced, more solid is fed to the raw sludge dewatering and the subsequent
digester sludge dewatering. The predicted total sludge solid treated could increase from 193.5
tonnes to 203.0 tonnes. As shown in Table 5-13, the total electricity consumption therefore
increases very slightly, by up to 0.3%. For the electricity production, the percentage increase
could be up to 18.8%. The electricity deficit could also be substantially reduced by up to 20.0%.
Table 5-13 The electricity balance of models with different Raw sludge storage time
Tank volume, m3
27248 21798 16349 10899 5450 0
Primary Settlement, kWh/d 1217 1217 1217 1217 1217 1217
Aeration, kWh/d 20203 20203 20203 20203 20203 20203
Fixed consumption, kWh/d 70279 70279 70279 70279 70279 70279
Sludge Treatment, kWh/d 6417 6446 6510 6583 6651 6735
Total Consumption, kWh/d 98116 98145 98209 98282 98350 98434
Electricity Generation, kWh/d 51352 52012 53281 54917 57251 61020

Electricity Deficit, kWh/d 46764 46132 44928 43365 41099 37414
The conclusion from this is that reducing the sludge storage time could prevent loss of energy-
containing material during sludge dewatering, and therefore enable greater energy recovery. In
the example of WWTP C, the raw sludge storage tank is built for contingency if the downstream
processes of THP or digester is shut down due to broken down or being maintained (personal
communication, L Wilkinson, Northumbrian Water, October 2018). The long sludge storage
time is not intentional but is due to the building up of sludge is because of the occasionally shut
down of certain THP and digester (personal communication, L Wilkinson, Northumbrian Water,
October 2018). Therefore, it is possible to reduce the sludge storage time.
5.4.3 Case study: Recovering energy from the dewatered digested sludge via pyrolysis
As shown in the previous Section 5.3.2 in WWTP C, approximately 20% of the sludge chemical
energy is stored in the dewatered digested sludge. This case study investigates the use of
pyrolysis to recover this energy for electricity production.
136
5.4.3.2 Method
This pyrolysis modelling is built on the model CM-4-1, in which a solid removal efficiency of
primary settlement of 0.4 is assumed. The TS content of the dewatered digested sludge is
estimated using Equation 5-7.
The design of the pyrolysis module is based on Mills et al. (2014). The module has three units:
1. sludge drying, which utilizes low-grade heat to dry the sludge to reach 90% DS (Mills et al.,
2014)
2. fast pyrolysis to convert biosolids into syngas
3. CHP for combusting the syngas to produce electricity and heat production.
A schematic illustration of the process is shown in Figure 5-20.
Biomethane
Boiler
CHP
Thermal Anaerobic
Hydrolysis Digestion
Sludge
Drying
Syngas
Pyrolysis
CHP
Sludge Biomethane/Syngas Electricity

High-grade heat Low-grade heat
Figure 5-20 Schematic illustration of the dewatered digested sludge pyrolysis module
137
The key assumptions made for this case study are as follows:
• In the sludge drying, 0.9 MWh heat is required per tonne of water evaporated (Mills et
al., 2014). 0.055 MWh electricity is required to process each tonne of TS content (Mills
et al., 2014). Both high-grade heat and low-grade heat can be used for the heat supply
• In the pyrolysis, 80% of the chemical energy stored in the sludge will be converted into
syngas (Mills et al., 2014, Cao and Pawłowski, 2012). Meanwhile, 0.4 MWh electricity
is required to treat each tonne of DS (Mills et al., 2014).
• In both CHP units, the electricity conversion ratio is 38%, and both the high grade heat
conversion ratio and low grade heat conversion ratio is 20% (Bowen et al., 2010, Smyth
et al., 2016).

When the solid removal efficiency of primary settlement is 0.4, conceptual model CM-4-1
predicts the dewatered digested sludge has a COD concentration of 257,417.2 mg/L and a
predicted flow of 78.9 m3/d. Via Equation 5-7, TS concentration is estimated as 265,655.8 mg/L.
Since the density of sludge is assumed to be 1,000,000 mg/L in this study (explanation shown
in Appendix C). The TS% of the dewatered digested sludge is estimated as 26.5%. The TS
loading is 21.0 tonnes. Via Equation 4-2, chemical energy loading of the dewatered digested
sludge is estimated as 87.1 MWh.
In the sludge drying process the electricity consumption is 19.0×0.055 = 1.155 MWh and the
heat consumption is 0.9×[(21.0÷26.5%)-(19.0÷90.0%)] = 52.3 MWh. For the pyrolysis,
0.4×21.0=8.4 MWh electricity is required. Therefore, the total energy consumption of the
module is 52.3 MWh heat and 9.6 MWh electricity.
For energy production, 87.1×80%=69.7 MWh syngas is produced. In the subsequent CHP
application, 69.7×38%=26.4 MWh electricity and 69.7×20%+69.7×20%=27.9 MWh heat
(combines the high-grade heat and low-grade heat) are produced. Therefore, the net electricity
production is 16.8 MWh but there is a heat deficit of 24.4 MWh. The predicted biomethane
production is 159.7 MWh in conceptual model CM-4-1, when the solid removal efficiency of
primary settlement of 0.4. The heat deficit will be supplemented by the low-grade heat
produced in the biomethane CHP since the production is 159.7 MWh×85%×20%=27.1 MWh.
138
As stated in Section 5.4.1, the electricity deficit of the WWTP is approximately 46.7 MWh.
With the net electricity output from the pyrolysis module is 16.8 MWh, the gap is reduced by
35.9% to 29.9 MWh.
Unlike the previous two case studies, in which the process modification is upstream of the
digestion, the pyrolysis targets the end product, and hence has the advantage that it would have
no impact on upstream processes or energy recovery of the digestion. The key to success of
this approach is to meet the low-grade heat demand for the sludge drying. Moreover, the two
case studies focused on operational adjustments, whereas the sludge pyrolysis would require
capital investment for installation of sludge drying, pyrolysis and the CHP infrastructure.
5.4.4 Case study: Replacing the activated sludge processes with an energy-generating
anaerobic membrane reactor unit
This case study investigates the benefits of replacement of the aerobic activated sludge
treatment process with an anaerobic membrane reactor (AnMBR) for reducing the energy
consumption of biological treatment of the wastewater.
5.4.4.2 Method
A conceptual model CM-4-2 is built on CM-4-1 in GPS-X (as shown in Figure 5-21). The
predicted result generated by the CM-4-2 will be used to construct an energy balance. The
newly obtained energy balance will be compared to the baseline case energy balance obtained
from the CM-4-1 while the SREPS is 0.4. There are several key adjustments made in the model
CM-4-2 comparing to CM-4-1:
• The “Anaerobic MBR” is used to replace the existing activated sludge process.
According to Gimenez et al. (2011) and Gouveia et al. (2015), the tank size is given
66,406 m3 for achieving the HRT to approximately 6 hours, and the mixed liquor
suspended solid (MLSS) is set as 20,000 mg/L and is control by a PID controller. The
stoichiometry and kinetics of the biological reactions of the “Anaerobic MBR” remain
at their default value in the model. The treated effluent of the AnMBR will be
discharged as final effluent.
• The solid removal efficiency of the primary settlement remains fixed at 0.40. The
139
surplus AnMBR sludge (SAnS) produced will be thickened in a similar way to the SAS
thickening in model CM-4-1. However, because the MLSS of the AnMBR is set as
20,000 mg/L instead of the 2,805 mg/L used in CM-4-1, the “fraction of influent flow”
of the SAnS drum thickener is changed to 0.35 in order to obtain thickened SAnS with
similar COD content of the thickened SAS (in the original model CM-4-1).
• The kinetics and model stoichiometry of the biological reactions of all of the sludge
storage tanks, and the rest of the thickening and dewatering facilities, will not be
changed.
• In anaerobic digestion, the fraction of “P content of the inert particulate material” of
the digester feed is changed from 0.021947 to 0.01794 in order to allow the feed sludge
to have the same extent of digestion as in the baseline case.
Figure 5-21 The conceptual model CM-4-2
For electricity consumption, the AnMBR is assumed to be energy self-sustaining and the
original activated sludge related electricity consumption, including the consumption on
aeration, RAS and SAS pumping will be ignored in the calculation
5.4.4.3 Result and discussion

The chemical energy balance of the baseline case (uses aerobic activated sludge as secondary
treatment) and the AnMBR model are listed in Table 5-14.
140
Table 5-14 The chemical energy balance of the baseline model outputs and the AnMBR model outputs
Baseline AnMBR
Final Effluent 13.7% 10.8%
Ends Gaseous Product 31.6% 45.8%
Primary Sludge 29.5% 29.5%
SAS/SAnS 25.3% 13.9%
Primary Sludge Filtrate 2.8% 2.8%

SAS Filtrate 5.7% 2.3%
Raw Sludge Centrate 8.7% 3.9%
Dewatered Raw Sludge (as Digester Feed) 34.9% 30.7%

CH4 19.6% 17.2%
Dewatered Digested Sludge 10.7% 9.3%
Digested Sludge Centrate 1.7% 1.5%
Using AnMBR results in a lower COD concentration in the final effluent, and hence the
percentage chemical energy in the final effluent is less (10.8%) than if activated sludge is used
(15.1%). Moreover, the percentage chemical energy captured in the SAnS is less than what is
captured in the SAS. Therefore, in the AnMBR model, less chemical energy is fed to sludge
treatment. Anaerobic treatment is reported to produce less sludge than aerobic treatment
(Gimenez et al., 2011, Gouveia et al., 2015). The predicted COD and TSS loading of the SAS
of base line model are 46.9 tonnes and 40.3 tonnes, respectively. The predicted COD and TSS
loading of the SAnS are approximately half of the SAS loading and are 25.8 tonnes and 25.0
tonnes, respectively. Regarding the TSS produced per g COD removed, the predicted result is
0.23g TSS/g CODRemoved which is in the reported range of 0.16-0.55g TSS g CODRemoved
(Gouveia et al., 2015). Modelling the use of AnMBR, 45.7% of the inflowing chemical energy
ends up as gaseous product. Unlike the model results for the activated sludge process, the
product is CH4 rather than CO2. The CH4 production is 22,710 m3 and results in a specific
methane yield of 0.21L CH4/g CODRemoved which is also in the reported range of 0.13-0.24 L
CH4/g CODRemoved (Gouveia et al., 2015, Gimenez et al., 2011). Although the model uses the
default kinetic and stoichiometry of biological reaction in the AnMBR model, the predicted
result is in agreement with the literature value.
During the sludge treatment prior to digestion, the major difference is in the raw sludge centrate.
The AnMBR process has a weaker raw sludge centrate because the biomass contained in the
141
SAnS has very little heterotrophic biomass and hence the solubilization of the particulate COD
is weaker. In the digestion, less CH4 is produced in the AnMBR model because the sludge
produced from the wastewater side is less, but the difference is small.
With respect to electricity, the generation from CH4 produced from the sludge of AnMBR is
approximately 12% less than if the activated sludge process (the baseline model) was used, as
shown in Table 5-15. However, this study disregards the electricity consumption of the original
activated sludge process, whilst the AnMBR is primarily assumed energy self-sustaining. In
this sense, the electricity consumption is reduced substantially by 33.9%, as is the electricity
deficit (by 57.7%).
Table 5-15 The electricity balance of the activated sludge and the AnMBR processes
Activated Sludge AnMBR
Primary Settlement, kWh/d 1217.1 1217.1
Aeration, kWh/d 20202.9 0.0
Fixed consumption, kWh/d 70278.9 58496.3
Sludge Treatment, kWh/d 6417.2 5143.6
Total Consumption, kWh/d 98116.1 64857.0
Electricity Generation, kWh/d 51352.2 45065.1

Electricity Deficit, kWh/d 46763.9 19791.9
The biggest uncertainty that affects the energy balance is whether the AnMBR process is able
to be energy self-sustaining. Martin et al. (2011) reported that the energy demand of the
submerged AnMBR varies from 0.03-3.57 kWh/m3. In this case study, the CH4 produced from
the AnMBR is 22,710m3×0.651kg/m3×55,500kJ/kg÷3600kJ/kWh=227,923 kWh. While the
flow of raw wastewater is 277,236 m3/day, the energy produced from 1 m3 wastewater is 0.82
kWh. Therefore, there is a possibility that the AnMBR is not energy self-sustaining. In the case
if the AnMBR requires 3.57 kWh energy to treat 1 m3 influent, 989,732 kWh heat energy is
required. The heat deficit of the AnMBR will be 761,709 kWh. The implementation of the
AnMBR process would have a high degree of uncertainty. Moreover, although full scale
application has been reported, the requirement for COD concentration of the feed wastewater
is measurable in tens of thousands (Basile et al., 2015); in all four investigated WWTPs the
COD of the raw wastewater is no greater than 1,200 mg/L(Table 5-1), and of the settled
wastewater is no greater than 700 mg/L (Table 4-1).
142
5.5 Chapter summary
This study constructed, calibrated and validated models for the wastewater and sludge
treatment processes of four WWTPs. In all cases the predicted results differed from the
observed results by no more than ± 15%. These predicted results were then used to construct
COD and chemical energy mass balances for the WWTPs. Finally, the potential to improve
the energy efficiency of the WWTPs was investigated by modelling the impact of various
operational and process changes to the WWTPs.
The mass and energy balance of the four WWTPs shows that:
• the primary settlement captured 46.3 ± 12.0% of influent COD and chemical energy in
the primary sludge. The remainder flowed into the secondary treatment.
• Activated sludge as the secondary biological treatment captured 44.4 ± 15.5% of its
own influent COD and chemical energy in the SAS. Meanwhile, 42.2 ± 13.0% of the
influent COD and chemical energy to the activated sludge is lost, potentially as CO2.
The remaining 13.6 ± 5.4% will leave the treatment works as final effluent.
• Overall, 69.7 ± 12.8% of COD and chemical energy flowing into the works can be
captured in the sludge and sent for sludge treatment via digestion, and the rest is lost,
either to atmosphere or in the final effluent.
In the sludge treatment, due to several stages of thickening and dewatering prior to the digestion,
only 65.3% of the influent sludge energy actually reaches the digester, on average. Since the
digester has a 57.0% of energy recovery efficiency, 38.8% of the chemical energy fed to the
sludge treatment can be recovered as CH4, whilst 20.4% of the sludge energy remains in the
dewatered digested sludge, and approximately 30% of the sludge energy is left in the filtrates
and centrates.
On this basis, approximately 20-30% of the chemical energy in the raw wastewater has the
potential to be recovered as CH4. However, only 6-9% of the chemical energy in the raw
wastewater can be potentially recovered as electricity. In this study, even though the chemical
energy within the raw wastewater of the investigated WWTPs is 5-14 times higher than its
electricity consumption, the WWTPs are not always electricity self-sustaining.
143
The results of this research suggest that it should be feasible to improve the energy balance at
WWTPs by:
• improving the efficiency of solids removal in primary settlement
• reducing the sludge retention time prior to the digestion
• recovering energy from the dewatered digested sludge via pyrolysis, and
• reducing the electricity consumption by replacing the activated sludge processes with
energy-generating anaerobic membrane reactor (AnMBR) units.
Modelling case studies were undertaken around these four areas, and results indicated that the
electricity deficit could be reduced by implementing operational and/or process changes.
However, some approaches have significant drawbacks and uncertainties. For example, the
coagulant used for improving the settling performance of the primary treatment may have
adverse effects on sludge digestion, and the energy balance of AnMBR units is highly uncertain.
The implementation in the real world needs further investigation.
144
Chapter 6 Results and Discussion: Nitrogen and Phosphorus Mass
Balances of Wastewater Treatment Plants
This chapter presents the nitrogen and phosphorus balances of each process of the four WWTPs
studied (Section 6.1). Based on these balances the nutrient recovery opportunities from
wastewater and sludge are discussed in Section 6.2. Case studies are also presented, which
investigate the potential to improve nutrient recovery from sludge at WWTPs. In Section 6.3
the impact of implementing energy recovery technologies on the effectiveness of nutrient
recovery is discussed.
6.1 Nitrogen and phosphorus mass balance of individual treatment processes
6.1.1 Nitrogen and phosphorus mass balances of primary settlement units
nitrogen and nitrogen and phosphorus mass balances of primary settlement units.
In the case of WWTP A and D which have reactive settler, the sum of the total nitrogen (TN)
and total phosphorus (TP) of the sludge and effluent wastewater from the primary settlement
are >99% of the inflowing. The amount of TN and TP cannot be accounted for is considered as
loss but is negligible i.e.<1.0%. Therefore, even there is the biological sludge is recirculated
back to the primary settlement, the TN and TP loss caused by the biological reaction can be
ignored. The difference of the inflowing and outflowing TN and TP is smaller than 1.0%, and
hence both the nitrogen and phosphorus mass balance is considered reliable. The mass balance
is shown in Figure 6-1.
In terms of the distribution pattern, after primary settlement, 8-22.8 % of inflowing TN and
16.1-50.5 % of inflowing TP is captured in the primary sludge and 76.7-92.0% of the inflowing
TN and 49.6-83.9% of inflowing TP remains in the primary effluent.
The TN and TP captured by the primary sludge is positively related to the COD captured by
the primary sludge (shown in Figure 5-5). The main removal mechanism in the primary
145
settlement tanks is physical settling of particulates (Tchobanoglous et al., 2003). The particulate
nitrogen and phosphorus settled and captured in the primary sludge is proportional to the COD
settled in primary sludge. In the case of WWTP A, the TP captured by the primary sludge is
substantially higher than in the other three treatment plants. That is because this site has ferric
dosing and hence part of the soluble phosphate is converted into particulate ferric phosphate
which then settles (Tchobanoglous et al., 2003).
Figure 6-1 The nitrogen and phosphorus mass balances of primary settlement in a daily basis at the four
WWTPs investigated
At the four study WWTPs the percentage of the influent TP captured in the primary sludge is
greater than the percentage of TN. That is likely because in the raw wastewater of each plant
the ratio of the estimated particulate phosphorus to TP is greater than the ratio of the estimated
particulate nitrogen to TN (consists of mainly the total Kjeldahl nitrogen (TKN)), as shown in
Table 6-1. Using WWTP C and D (which have no ferric dosing for phosphate removal) as
examples, particulate phosphorus is 36.2% of TP for WWTP C and 58.3% for WWTP D. This
is approximately twice the particulate nitrogen percentage of TN (16.3% for WWTP C and
21.8% for WWTP D).
146
Table 6-1 The particulate phosphorus, TP, the particulate nitrogen and TN of the raw wastewater of the four
study WWTPs
particulate TKN TKN, mg/L TN, mg/L particulate phosphorus TP
WWTP A 13.3 50.9 51.16 4.38 8.52
WWTP B 15.1 56.93 56.93 1.82 9.13
WWTP C 9.06 55.5 55.5 3.92 10.82
WWTP D 6.93 31.83 32.13 2.32 3.98
Literature values suggest that the percentage of nitrogen captured in primary sludge is only 10-
25 % (Kristensen et al., 2004, Sötemann et al., 2006). Therefore, 75-90 % of nitrogen in
wastewater treatment works flows to the secondary treatment process. For phosphorus removal,
if there is no chemical dosing, 75-80 % of the phosphorus to the treatment works will flow to
secondary treatment. If chemical dosing is used then 18-80% of the phosphorus mass entering
primary settlement will flow to secondary treatment (Kristensen et al., 2004, Grizzetti and
Bouraoui, 2006) and Grizzetti and Bouraoui (2006). The nitrogen and phosphorus mass
balances reported in this study are similar to these literature values.
6.1.2 Nitrogen and phosphorus mass balance of secondary biological treatment as activated
sludge
nitrogen and nitrogen and phosphorus mass balances of activated sludge.
In all four activated sludge systems the differences between inflowing and outflowing
phosphorus mass are negligible. The phosphorus mass balance is considered reliable. The
nitrogen mass balance of the activated sludge process of WWTP B and C are also considered
reliable because the inflowing nitrogen loading equals the outflowing. For WWTP A and D,
their nitrogen balances are not closed because part of nitrogen is missing from the sludge and
wastewater effluent suggested by the modelling results. The missing nitrogen loading is
assumed left as gaseous product to the atmosphere. The nitrogen and phosphorus mass of the
activated sludge process and its subsequent clarification is shown in Figure 6-2.
147
Figure 6-2 The nitrogen and phosphorus mass balances of secondary biological treatment in a daily basis at the
four WWTPs investigated
The percentages of TN and TP captured by the surplus activated sludge (SAS) are also
positively proportional to the COD captured by the SAS (Figure 5-12). However, the reason
for this relationship does not appear to be the same as in the primary settlement (the proportion
of TN and TP that is in particulate form). The proportion of nutrient captured in the SAS
appears mainly to be because of biomass assimilation. The biomass requires carbon, nitrogen
148
and phosphorus to grow (Droste, 1997, Tchobanoglous et al., 2003). The biomass grown is
generally expressed as C5H7O2NP0.074 (Droste, 1997), and the GPS-X model expresses it as
C5H7O2N0.8P0.1K0.02Mg0.02Ca0.01. Thus, if more biomass is produced, a greater percentage of
TOC (also COD) is captured in the SAS in the form of biomass, and the same applies to the
nitrogen and phosphorus.
This study found that a considerable amount of nitrogen is lost in the aeration and final
clarification units in WWTPs A and D, whereas this is not the case in WWTPs B and C.
Unintentional denitrification during aeration and final clarification has been reported
previously (Flores-Alsina et al., 2010, Ji et al., 2015, Satoh et al., 2003), and that may be the
reason for N loss in WWTPs A and D. For denitrification to occur there must be a good source
of nitrate and an anoxic environment (Tchobanoglous et al., 2003). WWTPs B and C do not
include a nitrification process. Hence no nitrate is produced and there is no denitrification.
WWTPs A and D do have nitrification, and hence nitrate is available and denitrification is
possible.
The anoxic environment is formed differently in the activated sludge tank than in the final
clarification. Denitrification occurs in the aerated activated sludge due to a low DO
concentration, which means that the inner core of the biomass floc remains in an anoxic
condition (Tchobanoglous et al., 2003, Ji et al., 2015). Wilson and Bouwer (1997) reported the
threshold DO that inhibits denitrification ranges from 0.08-7.9 mg/L. In this study, the DO
concentrations of WWTP A and WWTP D are 2.0 mg/L and 0.66 mg/L respectively. They both
smaller than the highest reported threshold DO of 7.9 mg/L. Therefore, denitrification could
take place. In the final clarification, since there is has no aeration involved, anoxic or anaerobic
condition is formed naturally (Flores-Alsina et al., 2010). This also potentially enables the
denitrification.
Denitrification requires not only nitrate but also a source of readily biodegradable substrate
(Tchobanoglous et al., 2003). The activated sludge tank is likely to have sufficient substrate for
more denitrification to occur because it is where the substrate rich primary effluent first flows.
But the final clarification, which is after the activated sludge process, may have less available
substrate. Thus, less denitrification is likely during final clarification than in the activated
sludge process.
149
The nitrogen loss can also be attributed to nitrous oxide (N2O), which is an intermediated
product of the denitrification process (Foley et al., 2010). However, because the gaseous
product was not collected in the sampling campaign, it was not possible to confirm the
production of N2O during this study.
Apart from causing potential nitrogen loss, the nitrification also affects the composition of the
nitrogen compounds in the activated sludge treated effluent. As shown in Table 6-2, for
WWTPs A and D with nitrification, the nitrate and nitrite nitrogen (NO3--N and NO2--N)
contributes 42.6-78.4% of the TN, whilst for WWTPs C and D, the ammonium nitrogen (NH4+-
N) contributed 85.5-93.9% of the TN.
Table 6-2 The predicted concentrations of different types of nitrogen of the activated sludge treated effluent at
the four WWTPs investigated
NH4+-N, mg/L NO2--N, mg/L NO3--N, mg/L TN, mg/L
WWTP A 0.9 2.4 13.7 20.5
WWTP B 31.9 0.0 0.0 33.9
WWTP C 38.0 0.0 0.0 44.4
WWTP D 3.0 0.6 4.7 12.2
6.1.3 Nitrogen and phosphorus mass balance of sludge thickening and dewatering
nitrogen and nitrogen and phosphorus mass balances of sludge thickening and dewatering.
The differences between inflowing and outflowing nitrogen and phosphorus mass of each
sludge thickening and dewatering facility were <0.5% and were negligible. Therefore, the mass
balances are considered closed. The nitrogen and phosphorus mass balances of each thickening
and dewatering processes in WWTP C are shown in Figure 6-3.
150
Figure 6-3 The nitrogen and phosphorus mass balances of the sludge thickening and dewatering processes in
WWTP C
The percentage TN, TP, TSS and VSS mass distributed in the filtrates or centrates and in the
thickened sludge are similar to each other in the drum thickening of primary sludge and SAS,
particularly in the case of the SAS thickening. In both cases, the numerical differences of the
percentage mass distributed in the filtrates and in the thickened sludge between the four
substances are relatively small. However, the difference becomes substantial in the dewatering
of the raw sludge and digested sludge. As shown Figure 6-3, in the raw centrifuge and digested
centrifuge, the percentage TN and TP mass distributed in the centrate were 188%-801% of the
percentage TSS and VSS mass distributed in the centrate. This is because the portion of the
particulate TN and TP supposed to be captured in the sludge is solubilized in to ammonium and
phosphate due to the hydrolysis and fermentation after long storage times or digestion, and
hence remains in the centrate (Bouzas et al., 2007, Münch and Barr, 2001). Therefore, the
composition of the TN and TP in different filtrates or centrates is highly varied, as shown in
Table 6-3. In the primary sludge filtrate and SAS filtrate, the soluble fraction accounts
commonly for no more than 60% of the total nutrient. But in the raw sludge centrate, the soluble
PO43-- P accounts for 55.5% of TP and soluble NH4+-N accounts 82.2% of the TN. In the
151
digested sludge centrate, the soluble PO43-- P accounts for 68.5% of the TP and the soluble
NH4+-N accounts 85.8% of the TN.
Table 6-3 Nutrient concentration of different filtrates or centrates
NH4+-N, mg/L TN, mg/L PO43-- P, mg/L TP
Primary sludge filtrate 164.6 273.9 31.7 200.3
SAS filtrate 40.6 88.5 5.7 34.4
Raw sludge centrate 491.3 597.8 178.7 322.0
Digested sludge centrate 2238.7 2608.7 228.5 333.7
6.1.4 Nitrogen and phosphorus mass balance of anaerobic digestion
nitrogen and nitrogen and phosphorus mass balances of anaerobic digestion.
After anaerobic digestion, the inflowing nitrogen and phosphorus remain in the digested sludge.
This matches the reported findings that almost all of the nutrient content remains in the
digestate (Shi, 2011, Rasi, 2009).
6.2 Discussion of the recovery opportunities based on plant wide mass balances
6.2.1 Nutrient product of wastewater treatment
The nitrogen and phosphorus balances of all four WWTPs were constructed from the result
predicted via modelling and are presented in Figure 6-4. The nitrogen and phosphorus missing
from the outflowing sludge and wastewater are considered loss.
As shown in Figure 6-4, WWTPs can be considered as reactors with three outputs: the gas, the
final effluent, the sludge. Generally, the recovered nutrient is for fertilizer use (Mayer et al.,
2016). To evaluate whether WWTP outputs are a product or not depends on whether the
nitrogen and phosphorus contained within the products of wastewater treatment are fertile and
whether there is potential to recover them.
If a WWTP includes a nitrification or denitrification process a portion of the nitrogen could

end in the output gas (Siegrist et al., 1995, Foley et al., 2010, Tchobanoglous et al., 2003). In
this study, WWTP A and WWTP D both possess a nitrification process. 17.9 % of the inflowing
TN to WWTP A ends in the gas, and for WWTP D is 33.9% is lost as gas. The main ingredient
152
of this nitrogen stream is expected to be nitrogen, although sometimes N2O will occur as well
(Kim, 2013, Flores-Alsina et al., 2010, Tchobanoglous et al., 2003). Although the nitrogen in
the gas output is considerable in WWTPs A and D, irrespective of whether it is present as N2
or N2O it will have no fertility (Mayer et al., 2016). Hence, for nitrogen recovery from
wastewater the gas output is not considered a product and may even be considered a loss since
the denitrification process converts the valuable ammonium into N2, which has no value.
The predicted phosphorus mass balance suggests that the phosphorus only exists in the
outflowing sludge and wastewater but not present in the gas output.
Figure 6-4 The nitrogen and phosphorus balances of the four WWTPs investigated
153
Unlike the COD mass balance and chemical energy balance, the percentage of influent nitrogen
and phosphorus left in the final effluent is considerable, especially for WWTP B and C which
do not have nitrification processes (more than 55% of the inflowing nitrogen remains in the
final effluent). Therefore the final effluent represents a good reserve of nutrient.
Nutrients can be recovered from secondary effluent via ion exchange or in an algal reactor.
Both technologies are able to recover the cation NH4+ and the anions PO43- and NO3-
simultaneously (Wang et al., 2010, Kim and Benjamin, 2004, Liberti et al., 1981).
For ion exchange, a resin or zeolite is first used to capture the nutrient ion. Resin or zeolite
have functional groups with electric charges. The resin or zeolite with a positive charge
captures anions, and those with negative charges capture the cations. When the functional
group of the resin or zeolite is completely replaced by the nutrient ion, they will are washed
with strong ionic strength solution, such as NaOH or H2SO4, to restore the ion exchange
capacity by removing the nutrient from the functional group (Deng, 2014). The volume of the
washing solution is commonly much less than the volume of the wastewater treated. Hence,
the concentration of nutrient is elevated and is suitable for fertilizer production, such as in the
form of Struvite (Deng, 2014).
Algal reactors can uptake nitrogen and phosphorus nutrients simultaneously if the correct algal
species are in place (Wang et al., 2010). After the algae is harvested, dewatered, and chemically
or thermally destroyed, the nutrient stored in the algal cell either ends up in the liquid product
or in the solid residue (Shakya et al., 2017, Cai et al., 2013). The liquid product has a high
concentration of NH4+ and PO43-. As with the wash solution from ion exchange, it is suitable
for struvite production (Shakya et al., 2017). The solid residue can be used as a bio-solid and
applied on land for agricultural usage (Cai et al., 2013).
As shown in Table 6-2, the nitrification affects the composition of ammonium, nitrate and
nitrite in activated sludge treated effluent. Although the three nitrogen compounds can all be
recovered from the wastewater directly via algal reactor or ion exchange, the production of
nitrate and nitrite may cause nitrogen loss. In the case of WWTP A and D, part of the nitrogen
is lost as nitrogen gas in the activated sludge and subsequent final clarification processes, likely
due to unintentional denitrification (Foley et al., 2010, Siegrist et al., 1995, Ji et al., 2015). But
154
WWTPs B and C, which have no denitrification, do not suffer such losses. Therefore, removing
the nitrification process might prevent the nitrogen loss. Moreover, in the modelling of WWTP
B, changing the adjusted growth rate of ammonium oxidizers from 0.91 (calibrated value) to
1.20 enables nitrification, and the ammonium concentration changes from the original
predicted value of 31.9 mg/L to just 2.6 mg/L. Modelling results therefore suggest that this
amendment to the treatment process increases the predicted airflow to the activated sludge tank
from 152,400 m3/d to 331,500 m3/d, because nitrification requires extra oxygen to oxidize the
ammonium to nitrate. Consequently, the electricity consumption for aeration would likely be
increased. Removing the nitrification process not only prevents nitrogen loss, allowing greater
amounts of nutrient available for potential recovery, but also saves energy.
However, such ion exchange and algal reactor technologies are not widely applied. For ion
exchange, it has not yet been applied at full scale, and its adaptability to variable flows regime
is uncertain (Bunce et al., 2018). For algal reactors the bottleneck is their long HRT and high
temperature requirement (Wang et al., 2010). HRT values measurable in days require much
larger footprints than existing treatment processes which require HRT values of only a matter
of hours, or up to a day. Moreover, the optimal algal growth rate requires high temperatures
and extra heat input in cold climates (Wang et al., 2010).
The remaining output of the wastewater treatment is the sludge. 22.2-43.8% of nitrogen and
53.2-84.7% of the phosphorus flowing into the works are captured in the sludge in the four
studied WWTPs (shown in Figure 6-4). The sludge produced can be a product as it can be
applied on land directly (DEFRA, 2018), or it can be deemed as an intermediate product
because in some cases the sludge is subjected to anaerobic digestion before the final product is
utilized (Münch and Barr, 2001).
6.2.2 Nutrient product of sludge treatment
A mass balance for the sludge treatment was only investigated for treatment plant WWTP C.
This mass balance constructed from the predicted result obtained via modelling is shown in
Figure 6-5.
155
Figure 6-5 The nutrient balance for the sludge treatment process of WWTP C
(The value shows in the () are the percentage mass balance using the TN and TP inputted to the sludge treatment
as reference)
The TN and TP input to the sludge treatment process are from the indigenous sludge, imported
sludge and imported dewatered sludge, and are equivalent to 37.1% of the TN and 66.3% of
the TP flowing into WWTP C. As shown in Figure 6-5, during sludge treatment 68.5% of the
sludge TN and 64.5% of the sludge TP ends up in the filtrates and centrates, and 31.5% of the
sludge TN and 35.5 % of the sludge TP remain in the dewatered digested sludge. Similar same
proportions of TN and TP are found end up in filtrates and centrates and digested sludge
respectively. Shi (2011) studied the nutrient mass balance of a sludge treatment only has sludge
thickening of SAS and digested sludge and a conventional sludge anaerobic digestion. It
suggested that 50.4 % of the sludge TN and 47.5 % of the sludge TP ends up in the filtrates and
centrates, and the rest the sludge TN and TP remain in the dewatered digested sludge. It is
likely that more sludge thickening and dewatering process will increase the proportion of
nutrient distributed in the filtrates and centrates, though this is based on two studies and would
therefore need further investigation.
156
Although energy and nutrient recovery occur in the same sludge treatment process, the
definition of product varies in the two recovery processes. Biogas is categorized as a product
from an energy recovery perspective. However, biogas contains very little nutrient, which
largely remains in the filtrates, centrates and dewatered digested sludge (as shown in Figure 6-
5). The filtrates and centrates can be processed for nutrient recovery and the sludge is already
a nutrient product (DEFRA, 2018, Münch and Barr, 2001). Therefore, the filtrates and centrates
are considered an intermediate product and the dewatered digested sludge is categorized as a
product.
6.2.3 Nutrient recovery potential from wastewater
Since the dewatered digested sludge contains appreciable amounts of nitrogen and phosphorus
it can be applied to land as fertilizer (DEFRA, 2018). Therefore, it is still a product. In 2010,
approximately 80% of the UK sewage sludge was reused on land (DEFRA, 2012). As shown
Figure 6-5, 31.5% of the sludge TN and 35.5 % of the sludge TP remain in the dewatered
digested sludge in the sludge treatment process of WWTP C. Although WWTPs A, B and D
has no sludge digestion treatment, assuming their sludge produced from the wastewater
treatment will share the same distribution pattern as found in the sludge treatment of WWTP
C, the percentage TN and TP could be kept in their dewatered digested sludge are estimated
empirically and shown in Table 6-4
Table 6-4 The estimated percentage TN and TP kept in the dewatered digested sludge
%TN captured in %TP captured in
Estimated %TN Estimated %TP
sludge from sludge from
ends in digested ends in digested
wastewater wastewater
sludge** sludge***
treatment* treatment*
WWTP A 43.2% 13.6% 84.7% 30.1%
WWTP B 43.8% 13.8% 75.0% 26.6%
WWTP C 22.2% 7.0% 53.2% 18.9%
WWTP D 53.2% 16.8% 60.0% 21.3%
*: Collected from Figure 6-4
**: Estimated TN ends in digested sludge = TN captured in sludge from wastewater treatment ×31.5%
**: Estimated TP ends in digested sludge = TP captured in sludge from wastewater treatment ×35.5%
The result shown in Table 6-4 suggested that approximately 7-17 % of the TN and 19-30% of
the TP inflowing to the four WWTPs could be kept in the dewatered digested sludge. The result
also suggested that the dewatered digested sludge retain greater percentage TP inflowing to the
WWTP than TN.
157
Besides applying the dewatered digested sludge on land, nutrient recovery from sludge can be
via the most well-known struvite process, which converts NH4+-N and PO43+-P into
Mg(NH4)PO4 with the introduction of Mg2+ in a suitable pH range of 8-10 (Nelson et al., 2003).
Mg(NH4)PO4 is a slow release fertilizer and its industrial-scale commercial production from
sludge return liquor is proven (Ostara, 2010). For a WWTP with a more stringent nutrient
consent, the additional nutrient loading (due to the recirculation of nutrient rich filtrates and/or
centrates back to head of the work) might cause problems. Moreover, it may increase the
operational cost since, for example, more oxygen would be needed for the nitrification of the
ammonium and hence there would energy consumption for aeration (see Section 6.2.1). Also,
extra chemicals (e.g.iron or aluminum salts) may be needed for removal of the additional
phosphate (Ivanov et al., 2009). Nutrient recovery from sludge filtrates and/or centrates would
have two benefits: (1) recovery of valuable nitrogen and phosphorus, and (2) relieving pressure
on the wastewater treatment process, especially where a stringent nutrient effluent consent is
enforced (Jaffer et al., 2002, Münch and Barr, 2001).
Although both dewatered digested sludge and struvite are products of nutrient recovery, the
nutrient contained per unit weight of the product is different. The dewatered digested sludge
collected in the sampling campaign of this research has a TS% of 28.4%, a TN/TS% of 5.2%
and a TP/TS% of 1.8% (shown in Table 4-6). A gram of dewatered digested sludge contains
28.4%×5.2%×1g=0.015 g nitrogen and 28.4%×1.8%×1g=0.005g phosphorus. A according to
the chemical composition of struvite (Mg(NH4)PO4·6H2O), a gram of struvite contains 0.057g
nitrogen and 0.126g phosphorus. The nutrient content in the struvite is denser than the
dewatered sludge, which could save the haulage of the product. Moreover, struvite is not only
a fertiliser, but is also a chemical product for which there are industrial uses, such as the
production of fire-resistant panels and cement (Mayer et al., 2016).
The struvite process is viable if at least 20 mg/L NH4+ (15.6 mg/L NH4+-N) and 106 mg/L PO43-
(34.6 mg/L PO43--P) are present simultaneously (Stratful et al., 2001). From the data shown in
Table 6-3 and Figure 6-5, among the four filtrates and centrates, raw sludge centrate and
digested sludge centrate are the ideal reserves for recovery because (1) >40% of the sludge TN
and TP are retained in these centrates and (2) the NH4+-N and PO43--P concentration of both
centrates are typically hundreds or even thousands of mg/L. In contrast the NH4+-N and PO43—
P concentrations of the primary sludge filtrate and SAS filtrate are only just enough to meet the
158
minimum requirement for the struvite process (shown in Table 6-3). Struvite recovery therefore
commonly targets the digested sludge centrate (Münch and Barr, 2001). Nevertheless, as shown
in Figures 6-5, digested sludge centrate has the highest TN loading, and raw sludge centrate
has the highest TP loading.
From the data shown in Table 6-3 and Figure 6-5,
• 20.4% of sludge TN is kept in the raw sludge centrate and 82.2% of it is NH4+-N, then
the NH4+-N in the raw sludge centrate accounts for 20.4%×82.2% =16.8% of sludge
TN feed to the sludge treatment
• 31.6% of sludge TP is kept in the raw sludge centrate and 55.5% of it is PO43--P, then
the PO43--P accounts in the raw sludge centrate for 31.6%×55.5% =17.5% of sludge TP
feed to the sludge treatment
• 34.3% of sludge TN is kept in the digested sludge centrate and 85.8% of it is NH4+-N,
then the NH4+-N of the digested sludge centrate accounts for 34.3%×85.8% =29.4% of
sludge TN feed to the sludge treatment
• 12.6% of sludge TP is kept in the digested sludge centrate and 68.5% of it is PO43--P,
then the PO43--P of the digested sludge centrate accounts for 12.6%×68.5% =8.6% of
sludge TP feed to the sludge treatment
On this basis, similar to the approach of estimating the percentage TN and TP kept in the
dewatered digested sludge, the percentage of TN and TP inflowing to the work and existing as
NH4+-N and PO43-_P in the centrate can be estimated. The estimations of four WWTPs are
shown in Table 6-5.
Table 6-5 The estimated percentage TN and TP (existing as NH4+-N and PO43-_P) kept in the raw sludge centrate
and digested sludge centrate
%TN as
%TN as %TN as T%P as %TP as PO43-
NH4+-N in %TP as PO43-
NH4+-N NH4+-N in PO43--P in -P in RSC
RSC and -P in RSC
in RSC1 DSC2 DSC1 and DSC1
DSC
WWTP A 7.2% 12.7% 20.0% 14.9% 7.3% 22.2%
WWTP B 7.3% 12.9% 20.2% 13.2% 6.5% 19.6%
WWTP C 3.7% 6.5% 10.3% 9.3% 4.6% 13.9%
WWTP D 8.9% 15.7% 24.6% 10.5% 5.2% 15.7%
1: RSC =Raw sludge centrate
2: DSC = Digested sludge centrate
159
The result shown in Table 6-5 suggested that approximately 10-20 % of the TN and 14-22% of
the TP inflowing to the four WWTPs could exist as NH4+-N and PO43-_P and be kept in the raw
sludge centrate and digested sludge centrate for potential struvite recovery.
According to the chemical composition of Mg(NH4)PO4, the mass ratio between the nitrogen
(molar mass=14) and phosphorus (molar mass=31) is 1:2.2. Thus, not all the available TN and
TP could be recovered into the struvite simultaneously. In some cases the remaining NH4+-N
will be treated using the Anammox process, which converts the NH4+-N back to non-fertile N2
gas (Desloover et al., 2011). However, further recovery can be achieved. Due to the instability
of NH4+-N in alkaline conditions and its thermal instability, the redundant ammonium can be
stripped from the sludge filtrate and/or centrate in the presence of heat and at a suitable pH
(Bonmatı and Flotats, 2003). After the NH3 gas is stripped off the liquid, it will contact the
acidic solution which converts NH3 back to NH4+. This s kept and concentrated in the solution.
The NH4+ rich solution can then be further processed for fertilizer production (Bonmatı and
Flotats, 2003, Deng, 2014).
Based on the data shown in Table 6-4 and 6-5, approximately 17-41% of TN and 33-52% of
TP inflowing to the work could be kept in the dewatered digested sludge for later land
application or ends as NH4+-N and PO43-_P in the raw sludge centrate and digested sludge
centrate for fertilizer recovery. Table 6-4 and 6-5 also suggest that the greater percentage of
nutrient captured in the sludge in the wastewater treatment leads to a greater nutrient recovery
potential, for instance, among four WWTPs, WWTP D captured the greatest percentage TN
inflowing to the work, this leads to WWTP D has the greatest nutrient recovery potential of the
dewatered digested sludge and of the raw sludge centrate and digested sludge centrate.
6.3 Case studies of nutrient recovery
As stated in section 6.2, nutrient recovery is current conducted in the sludge treatment, from
the sludge dewatering centrates and the dewatered digested sludge. Therefore, this study first
examines the potential impact of greater nutrient being captured from the sludge via chemical
dosing or enhanced biological process prior to it being fed to the sludge treatment (Section
6.3.1).
160
Moreover, implementation of a particular recovery technology may impact the overall material
mass balances of WWTPs. In Section 5.4, three case studies are presented in which options for
improving the energy balance and they are:
• Improving solids removal efficiency in the primary settlement (Section 5.4.1)
• Reducing the sludge retention time prior to digestion (Section 5.4.2) and
• Replacing the activated sludge processes with an energy-generating anaerobic
membrane reactor (Section 5.4.4).
Since these three case studies are found will impact both the wastewater and the sludge COD
mass balance, Section 6.3.2 to 6.3.4 study their potential impact to the nutrient recovery.
The key criteria used to evaluate each of the improved energy recovery options case studies is
the change in NH4+-N and PO43--P loading in the raw sludge centrate and digested sludge
centrate where the struvite recovery could be implemented. Also, the change of nutrient mass
distribution pattern between the final effluent and digested dewatering sludge will be examined
for identifying the potent nutrient reserves.
6.3.1 Case study: Benefits to nutrient recovery of enhancing nutrient removal during
wastewater treatment
This study aims to capture more nutrient into the sludge. This study mainly focuses on the
phosphorus since applications of the chemical dosing in primary settlement or an Enhance
Biological Phosphorus Removal (EBPR) activated sludge process can improve capturing of
phosphorus into the sludge (Tchobanoglous et al., 2003).
6.3.1.1 Method
A conceptual model named CM-5-1 is developed from the conceptual model CM-4-1 (Section
5.4.1.2, shown in Figure 5-19) which has a solid removal efficiency (of primary settlement) of
0.4. A chemical dosing unit using ferric sulphate is added for enabling the chemical phosphate
removal. Moreover, a virtual anaerobic “Plug Flow Tank” with a set DO of 0 mg/L is added in
front of the original “Plug Flow Tank” for enabling the EBPR process. The volume of the
original “Plug Flow Tank” is set as 99,609 m3 instead of the original volume of 33,203 m3.
161
Moreover, the maximum growth rate of ammonia oxidizer of the original “Plug Flow Tank” is
set to 0.001 to prevent nitrification. Other settings all remain the same. The model of the CM-
5-1 is shown in Figure 6-6.
Figure 6-6 The CM-5-1 model
Three scenarios are investigated in this case study. The variables adjusted are the chemical
dosing rate and the volume of the anaerobic tank.
• In the baseline scenario, the model used is called the baseline model and is with a zero
chemical dosing rate and the volume of the anaerobic tank is set as only 1 m3.
• In the second scenario, the model used is called the ferric dosing model and is with
8,900 kg/d of ferric sulphate dosing and the volume of the anaerobic tank is set as only
1 m3. Hence, the extra phosphorus captured in the primary sludge is purely from the
chemical dosing.
• In the third scenario, the model used is called the EBPR model and is with a zero
chemical dosing rate and the volume of the anaerobic tank is set as only 33,203 m3. The
extra phosphorus captured in the sludge is all contained the biological sludge.
The predicted flow and concentrations are then used to demonstrate the influence to nutrient
mass balance of the system.
162
The modelling result shows that the Ferric dosing and EBPR are both able to capture more
phosphorus in the sludge in the wastewater treatment process (as shown in Table 6-6). In both
cases, approximately 75-80% of the phosphorus to the WWTP is captured in the sludge whilst
the baseline model only captures approximately 46%. Meanwhile, there is no substantial
impact to the nitrogen as the percentage of nitrogen captured in the sludge remains at
approximately 21% in all three cases.
Table 6-6 The TN and TP percentage mass balance of the baseline, ferric dosing and EBPR models
Baseline Model Ferric dosing model EBPR model
Percentage TN in final effluent 80.4% 79.0% 79.2%
Percentage TN loss in biological treatment 0.0% 0.0% 0.0%
Percentage TN in primary sludge 7.5% 7.7% 7.5%
Percentage TN in SAS 12.1% 13.2% 13.3%
Percentage TP in final effluent 54.0% 20.7% 24.4%
Percentage TP loss in biological treatment 0.0% 0.0% 0.0%
Percentage TP in primary sludge 15.1% 29.9% 15.1%
Percentage TP in SAS 31.0% 49.5% 60.6%
Although chemical dosing and EBPR capture similar amounts of phosphorus in the sludge,
they impact the phosphate content in the raw sludge centrate differently. As shown in Table 6-
7, the chemical dosing does not have a substantial impact nutrient content on the raw sludge
centrate or digested sludge centrate. This is because the phosphorus removed is present as
phosphate. This is bonded with the ferric iron as ferric phosphate which will rarely decomposed
during hydrolysis, fermentation and digestion. But for EBPR, the captured phosphorus is stored
in poly-phosphate accumulative organisms (PAOs). It will be released back to the liquid phase
as phosphate in an anaerobic condition, like in the raw sludge tank. Hence, the PO43--P content
of the raw sludge centrate of the EBPR model is about a third higher than the baseline model
and the model with Ferric dosing.
163
Table 6-7 The concentration and loading of NH4+-N and PO43--P in the raw sludge centrate and digested sludge
centrate of models with different phosphate removal scenarios
PO43--P in
NH4+-N in PO43--P NH4+-N in PO43--P in NH4+-N in
both
Model RSC1, in RSC1, DSC2, DSC2, both centrate,
centrate,
mg/L mg/L mg/L mg/L kg/d
kg/d
Baseline 539.8 212.5 1235.2 177.9 1191.6 326.7
With Chemical
514.5 190.7 1330.2 189.5 1298.9 337.4
Dosing
With EBPR 410.9 324.7 1168.9 155.4 1055.4 471.6
1: RSC stands for Raw sludge centrate
2: DSC stands for digested sludge centrate
The chemical dosing and EBPR have an impact on the TP distribution on the output of the
process but do not substantially influence the TN distribution as shown in Tables 6-8 and 6-9.
Noticeably, a substantial increase in TP loading is shown in the SAS filtrate in both cases. For
the chemical dosing, the increase of TP loading in the SAS filtrate and other filtrates and
centrates is because solids retained in the filtrates and centrates contain ferric phosphate (the
predicted concentration was 740.9 mgFePO4/L), but this is not present in the baseline model
nor the EBPR model. For EBPR, the TP loading increase is because the solid kept in the SAS
filtrate contains PAOs which is rich in phosphorus. The predicted concentration of poly-
phosphate in PAOs is 107.6 mg/L, but this is not present in the baseline model nor the ferric
dosing model. However, in both cases, the PO43--P content of the SAS filtrate is still no greater
than 10 mg/L.
Table 6-8 The TN distribution in the output of the combined wastewater and sludge treatment in models with
different phosphate removal scenarios
Primary SAS Raw Sludge Digested Sludge Dewatered
Model
Sludge Filtrate Filtrate Centrate Centrate Digested sludge
Baseline Model 0.5% 2.5% 5.0% 6.7% 4.8%
Ferric dosing
0.5% 2.8% 5.4% 7.0% 5.3%
model
EBPR model 0.5% 4.0% 4.5% 7.0% 4.8%
Table 6-9 The TP distribution in the output of the combined wastewater and sludge treatment in models with
different phosphate removal scenarios
Primary SAS Raw Sludge Digested Sludge Dewatered
Model
Sludge Filtrate Filtrate Centrate Centrate Digested sludge
Baseline Model 3.7% 5.0% 14.3% 5.5% 17.5%
Ferric dosing 5.7% 9.2% 21.2% 10.9% 32.4%
model
EBPR model 3.7% 24.7% 20.8% 6.1% 20.3%
164
It is also worth noting that the EBPR does not increase the TP loading in the dewatered digested
sludge because considerable amounts of it has entered into the filtrates and centrates. But
chemical dosing does increase the TP loading in the dewatered digested sludge as the ferric
phosphate is stable during the biological reaction and thus it remains in the solid fraction
throughout the process (Mayer et al., 2016). It therefore appears as though chemical dosing
will not boost the struvite recovery but will result in greater capture of influent TP in the
digested dewatered sludge. In contrast, the EBPR model predicts that slightly higher percentage
of TP would be captured in the digested dewatered sludge, but struvite recovery may be boosted
in the EBPR process.
In energy wise, as shown Table 6-10 and 6-11, the EBPR and chemical dosing have very mild
impact on the chemical energy and the electricity balance.
Table 6-10 The chemical energy balance of the models with different phosphate removal scenarios
Baseline Model EBPR model Ferric dosing model
Final Effluent 12.7% 12.3% 11.7%
Loss in Activated sludge 39.9% 38.9% 39.2%
Primary Sludge 29.5% 29.5% 29.8%
SAS 17.9% 19.3% 19.3%
Primary Sludge Filtrate 2.8% 2.8% 2.8%
SAS Filtrate 2.2% 3.9% 2.3%
Raw Sludge Centrate 7.5% 5.5% 7.9%
Dewatered Raw Sludge (as Digester Feed) 32.7% 33.4% 33.6%
CH4 18.3% 18.7% 18.9%
Dewatered Digested Sludge 10.0% 10.2% 10.3%
Digested Sludge Centrate 1.6% 1.6% 1.6%
165
Table 6-11 The electricity balance of the models with different phosphate removal scenarios
Baseline Model EBPR model Ferric dosing model
Primary Settlement, kWh/d 1217 1217 1230
Aeration, kWh/d 25393 22197 24958
Fixed consumption, kWh/d 70279 70279 70279
Sludge Treatment, kWh/d 6417 5775 5949
Total Consumption, kWh/d 103306 99468 102415
Electricity Generation, kWh/d 51352 48958 49393

Electricity Deficit, kWh/d 51954 50510 53023
6.3.2 Case study: Impact on nutrient recovery of improved primary settlement
Struvite is recovered from the sludge centrates, which arises from the combined primary sludge
and the SAS. A change in solid removal efficiency would influence the production of primary
sludge and SAS, and hence may also influence how much phosphate and ammonium is
available in the return liquor.
6.3.2.2 Method
The flow and concentration data predicted in conceptual modelling conducted in the Case
Study: Improving solids removal efficiency in primary settlement (Section 5.4.1) are used to
study the influence of improving solids removal efficiency in primary settlement on the nutrient
mass balance.

An increase in solid removal efficiency in primary settlement (SREPS) will increase the
percentage of nutrient captured in the primary sludge as more particulate nutrient settles.
However, less substrate COD will flow to the activated sludge process, and hence less nutrient
would be assimilated in the biological sludge. In this case study, a greater solid removal rate in
primary settlement will slightly reduce the percentage of nutrient captured in the sludge. As
shown in Table 6-12, when the SREPS is increased from 40% to 70% the percentage of influent
TN and TP captured in the combined sludge is only reduced from 22% to 20% and 54% to 46%,
respectively. Simultaneously the concentration of ammonium and phosphate of both centrates
166
reduces substantially, as shown in Table 6-6. With the increase of SREPS from 0.4 to 0.7,
although the predicted flow of the raw sludge increases from 1,264.1m3 to 1393.7 m3 and the
predicted flow of digested sludge increases from 395.3 m3 to 502.6 m3, the loading of both
NH4+-N and PO43--P kept in the two centrates reduces.
Table 6-12 The concentration and loading NH4+-N and PO43--P of raw sludge centrate and digested sludge
centrate under different solid removal efficiency of primary settlement
NH4+-N in NH4+-N in NH4+-N in PO43--P in

PO43--P in PO43--P in
SREPS RSC1, DSC2, both centrate, both centrate,
1
RSC , mg/L DSC2, mg/L
mg/L mg/L kg/d kg/d
0.40 581.8 237.2 1352.6 499.9 1422.5 554.6
0.45 551.4 219.3 1252.4 430.7 1370.1 510.1
0.50 509.7 195.7 1153.0 357.4 1304.0 454.4
0.55 461.5 169.3 1059.6 285.1 1229.1 392.0
0.60 410.5 142.7 974.0 217.0 1149.0 326.9
0.65 358.0 116.9 898.3 154.8 1067.7 262.1
0.70 303.9 92.5 828.8 98.9 981.2 198.3
As stated in Section 5.4.1.3, less heterotrophic biomass is produced in the activated sludge
process while the SREPS increases. This will leads to less fermentation and hydrolysis take
place in the raw sludge tank because the modelling of extent of fermentation and hydrolysis is
positively related to the content of heterotrophic biomass (Henze et al., 1999). It is likely the
cause of the reduction of the ammonium and phosphate concentration and loading in the raw
sludge centrate.
The cause of the reduction of the ammonium and phosphate in the dewatered digested sludge
centrate is likely because of the ratio of nitrogen and phosphorus contained in per unit of COD
reduces. Although more nutrient is fed to the digestion process, the amplification of it is smaller
than amplification of the COD. When the SRE increases from 0.4 to 0.7, the percentage of
influent COD content retained in the dewatered raw sludge increases by 32.6%, but the
percentage of influent TN retained in the dewatered raw sludge only increases by only 13.6%,
and TP retained in the dewatered raw sludge decrease by 8.5%. Since the conversion rate of
COD to biomethane constantly stays at approximately 50%, less nitrogen and phosphorus is
167
solubilized when the solid removal efficiency increases. Consequently, even more nutrient is
kept in the digester feed, and less ammonium and phosphate end in the digested centrate.
A greater SRE in the primary settlement process leads to a greater percentage of nutrient in the
final effluent and digested sludge cake, but less ammonium and phosphate in the raw sludge
and digested centrate available for struvite recovery. Improving the SRE of the primary
settlement aims to elevate the CH4 production. However, if struvite is a target product, then its
production will be compromised. The TN and TP distribution in the output of the combined
wastewater and sludge treatment is shown in Tables 6-13 and 6-14.
Table 6-13 The TN distribution in the output of the combined wastewater and sludge treatment under different
solid removal efficiency of primary settlement
Primary Dewatered
Final SAS Raw Sludge Digested Sludge
SREPS Sludge Digested
Effluent Filtrate Centrate Centrate
Filtrate sludge
0.40 77.6% 0.5% 3.4% 6.0% 5.5% 7.0%
0.45 77.9% 0.6% 3.0% 5.7% 5.4% 7.3%
0.50 78.3% 0.6% 2.7% 5.4% 5.3% 7.7%
0.55 78.7% 0.7% 2.4% 5.0% 5.2% 8.0%
0.60 79.1% 0.8% 2.1% 4.5% 5.2% 8.4%
0.65 79.5% 0.8% 1.8% 4.1% 5.1% 8.7%
0.70 79.8% 0.9% 1.5% 3.6% 5.1% 9.1%
Table 6-14 The TP distribution in the output of the combined wastewater and sludge treatment under different
solid removal efficiency of primary settlement
Primary Dewatered
Final SAS Raw sludge Digested Sludge
SREPS Sludge Digested
Effluent Filtrate Centrate Centrate
Filtrate Sludge
0.40 46.3% 3.7% 6.8% 17.2% 9.9% 16.1%
0.45 47.3% 4.2% 6.0% 16.6% 9.2% 16.6%
0.50 48.5% 4.7% 5.4% 15.9% 8.3% 17.2%
0.55 49.8% 5.1% 4.8% 15.0% 7.4% 17.9%
0.60 51.2% 5.6% 4.3% 14.1% 6.4% 18.5%
0.65 52.7% 6.0% 3.8% 13.2% 5.3% 19.0%
0.70 54.2% 6.5% 3.3% 12.2% 4.3% 19.5%
168
The uncertainty embedded in this case study is whether the increase of SRE will affect the
nutrient balance, especially the phosphorus balance. Literature suggests that SRE can be
improved by dosing with aluminium and ferric salts (Ismail et al., 2012, Sarparastzadeh et al.,
2007). However, both aluminum and ferric iron are able to react with phosphate. Thus, more
phosphorus is captured in the primary sludge, and it is also likely that less phosphorus will
remain in the final effluent if the SRE is raised by dosing with metal salt coagulants. However,
this will not affect the nitrogen balance since the nitrogen derivatives do not react with
aluminum and ferric ion.
6.3.3 Case study: Impact on nutrient recovery of reducing sludge storage time
The approach to reducing the particulate energy containing COD loss in the sludge dewatering
is to reduce the sludge storage time of the raw sludge. As stated in Section 5.4.2.3, less
particulate COD is solubilized by fermentation when the sludge storage time is reduced. Since
particulate nitrogen and phosphorus can also be solubilized into ammonium and phosphate
during sludge hydrolysis and fermentation (Tchobanoglous et al., 2003), reducing the sludge
storage time may affect the ammonium and phosphate content in the raw sludge centrate and
the digested sludge centrate.
6.3.3.2 Method
Study: Reducing the sludge retention time prior to digestion (Section 5.4.2) are used to study
the influence of the reducing the sludge retention time prior to digestion on the nutrient mass
balance

A smaller tank volume leads to a shorter sludge storage time. With a reduced sludge storage
time the extent of hydrolysis and fermentation in the raw sludge reduces. The ammonium and
phosphate produced from the hydrolysis and fermentation of the raw sludge also reduces. Since
more TN and TP remain and fed to the digester, the content of ammonium and phosphate of
the digested sludge centrate increases (as shown in Table 6-15). However, with sludge storage
time reduces, the loading of the ammonium and phosphate of the combined raw sludge centrate
and digested sludge centrate reduces.
169
centrate under different raw sludge tank volume scenarios
Tank NH4+-N in PO43--P in NH4+-N in PO43--P in
NH4+-N in both PO43--P both
volume, RSC1, RSC1, DSC2, DSC2,
centrate, kg/d centrate, kg/d
m3 mg/L mg/L mg/L mg/L
26086 581.7 237.2 1352.6 499.9 1422.4 554.5
20869 533.5 210.8 1385.4 521.7 1387.7 535.4
15652 473.9 177.5 1420.4 546.0 1344.8 510.8
10434 387.3 132.9 1474.3 576.3 1283.4 477.7
5217 254.2 75.8 1565.9 608.9 1193.5 434.1
0 41.1 6.1 1708.1 627.6 1053.9 375.5
The upstream process of the wastewater treatment and sludge thickening of primary sludge and
SAS stay the same. With the same input to the raw sludge dewatering and its subsequent
treatment, as less TP and TN are kept in the centrate, a greater proportion of nutrient will remain
in the dewatered digested sludge. Reducing the raw sludge storage tank volume aims to elevate
the CH4 production. However, if the struvite recovered from the raw sludge centrate and
digested sludge centrate is a target product, then its production will be compromised. The TN
and TP distribution in the output of the combined wastewater and sludge treatment are shown
in Tables 6-16 and 6-17.
Table 6-16 The TN distribution in the output of the combined wastewater and sludge treatment under different
raw sludge tank volume scenarios
Digested Dewatered
Tank Final Primary SAS Raw sludge
Sludge Digested
volume, m3 Effluent Sludge Filtrate Filtrate Centrate
Centrate Sludge
26086 77.6% 0.5% 3.4% 6.0% 5.5% 7.0%
20869 77.6% 0.5% 3.4% 5.6% 5.7% 7.2%
15652 77.6% 0.5% 3.4% 5.2% 5.9% 7.4%
10434 77.6% 0.5% 3.4% 4.5% 6.3% 7.7%
5217 77.6% 0.5% 3.4% 3.6% 6.9% 8.0%
0 77.6% 0.5% 3.4% 2.0% 7.9% 8.7%
170
Table 6-17 The TP distribution in the output of the combined wastewater and sludge treatment under different
raw sludge tank volume scenarios
Digested Dewatered
Tank Final Primary SAS Raw sludge
Sludge Digested
volume, m3 Effluent Sludge Filtrate Filtrate Centrate
Centrate Sludge
26086 46.3% 3.7% 6.8% 17.2% 9.9% 16.1%
20869 46.3% 3.7% 6.8% 16.3% 10.5% 16.4%
15652 46.3% 3.7% 6.8% 15.2% 11.1% 16.9%
10434 46.3% 3.7% 6.8% 13.7% 12.0% 17.5%
5217 46.3% 3.7% 6.8% 11.7% 13.1% 18.4%
0 46.3% 3.7% 6.8% 9.3% 14.3% 19.6%
6.3.4 Case study: Impact on nutrient recovery of using an anaerobic membrane reactor as
secondary biological treatment
In Section 5.4.4, anaerobic membrane reactor (AnMBR) treatment was found to produce less
sludge compared to an activated sludge process. Moreover, anaerobic treatment is normally not
efficient for nutrient removal (Chernicharo, 2006). Therefore, the nutrient distribution pattern
in the wastewater treatment may change dramatically if AnMBR is substituted for the activated
sludge process.
6.3.3.2 Method
Study: Replacing the activated sludge processes with an energy-generating anaerobic
membrane reactor unit (Section 5.4.4) are used to study the influence of replacing the activated
sludge processes with an energy-generating anaerobic membrane reactor unit on the nutrient
mass balance.

Compared to the baseline model that uses activated sludge, the application of AnMBR
substantial reduces the amount of NH4+-N and PO43--P in the centrates (as shown in Table 6-
18). This can be attributed to the sludge production of the AnMBR being much smaller than
the activated sludge model used by the baseline model, and hence less nutrient is captured in
the sludge that is fed to the sludge treatment. Moreover, since the AnMBR does not produce
the same quantity of heterotrophic biomass as the activated sludge process, the extent of
171
fermentation in the raw sludge tank is far smaller than the activated sludge process, and hence
content of NH4+-N and PO43--P of the raw sludge filtrate reduces substantially.
centrate of models with different secondary biological treatment
PO43--P in
NH4+-N in PO43--P NH4+-N in PO43--P in NH4+-N in
both
Model RSC1, in RSC1, DSC2, DSC2, both centrate,
centrate,
mg/L mg/L mg/L mg/L kg/d
kg/d
Activated Sludge 581.8 237.2 1352.6 499.9 1422.5 554.6
AnMBR 152.3 22.6 1281.4 19.1 715.0 31.9
The application of AnMBR also impacts on the TN and TP distribution in the outputs from the
treatment process (shown in Table 6-19 and 6-20). However, the impact is greater for TP. This
is because in the baseline model 38.6% of the TP to the WWTP is captured in the biological
sludge. But the figure is reduced to 9.7%% in the AnMBR case. This leads to the percentage
of TP captured in the sludge reducing from 53.7% to 24.8%. Meanwhile the percentage of TN
captured in the sludge only reduces from 22.4% to 15.9%. The most noticeable change in the
TP distribution pattern of the application of AnMBR is that the percentage of influent TP kept
in the final effluent increases from 46.3% to 75.2%. If struvite recovered from the raw sludge
centrate and digested sludge centrate is a target product, applying AnMBR may improve the
energy balance (stated in Section 5.4.4.3) but will compromise the struvite production. In
addition, more nutrient will remain in the final effluent.
Table 6-19 The TN distribution in the output of the combined wastewater and sludge treatment in different
scenarios of secondary biological treatment
Digested Dewatered
Final Primary SAS Raw Sludge
Model Sludge Digested
Effluent Sludge Filtrate Filtrate Centrate
Centrate Sludge
Baseline 77.6% 0.5% 3.4% 6.0% 5.5% 7.0%
AnMBR 84.1% 0.5% 2.1% 2.6% 4.6% 6.1%
172
Table 6-20 The TP distribution in the output of the combined wastewater and sludge treatment of models with
different secondary biological treatment
Digested Dewatered
Final Primary SAS Raw sludge
Model Sludge Digested
Effluent Sludge Filtrate Filtrate Centrate
Centrate Sludge
Baseline 46.3% 3.7% 6.8% 17.2% 9.9% 16.1%
AnMBR 75.2% 3.7% 2.9% 5.8% 1.6% 10.8%
6.4 Chapter Summary
This chapter first presents the nitrogen and phosphorus mass balance of the studied WWTPs.
In primary settlement, 8-23% of the inflowing nitrogen and 16-50% of inflowing phosphorus
will be settled in the primary sludge and the rest will flow into the secondary treatment. Since
the studied WWTP treatment works have no enhanced biological phosphorus removal
technology, 32-60% of phosphorus flowing to the secondary treatment is stored in the SAS and
the rest will leave the treatment works within the final effluent. Meanwhile, approximately 15-
30% of the nitrogen flowing to the secondary treatment is captured in the SAS. Where
nitrification is occurring, this will affect the distribution pattern of the remaining nitrogen
content. If there is no nitrification the remaining nitrogen will leave the system and is present
mainly as NH4+-N. For the WWTPs with nitrification, only approximately 50% of the nitrogen
flowing to the secondary treatment process remains in the effluent, mainly as nitrate or nitrite.
The rest is denitrified and emitted to the atmosphere due to unexpected denitrification occurring
in the aeration tank and in the final clarifier. After the whole treatment process, approximately
22-44% of the nitrogen and 53-85% of phosphorus is captured in the sludge. The rest of the
phosphorus will leave the WWTPs in the final effluent. For nitrogen, if nitrification does not
occur the remaining 55-78% of nitrogen will leave the WWTPs within the final effluent. Where
nitrification does occur, 18-34% of the nitrogen will be emitted to atmosphere and the rest
(approximately 40%) will leave the works in the final effluent. The distribution pattern of the
nitrogen and phosphorus are substantially different during the wastewater treatment process:
the majority of the phosphorus is captured in the sludge, but the majority of the nitrogen is
either emitted to the atmosphere or is discharged in the final effluent.
During sludge treatment with sludge thickening, dewatering and anaerobic digestion, the
distribution patterns of nitrogen and phosphorus are similar. Approximately 30-35% of the
sludge nitrogen and phosphorus remain in the dewatered digested sludge, but no nitrogen or
173
phosphorus is present in the biogas. The rest of the nitrogen and phosphorus is found in the
filtrates and centrates of the sludge thickening and dewatering. A large proportion of the
nitrogen and phosphorus, in the form of NH4+-N and PO43--P, are typically found in the filtrate
and/or centrate of the biological treated sludge, or sludge that has been stored for a long time.
This is because the particulate nutrient content in the sludge is solubilized and therefore enters
into the liquid phase.
Regarding nutrient recovery, although the final effluent contains substantial amounts of
nitrogen, its recovery is focused mainly on the sludge treatment side. Current practice is to
recover nutrients by either by (1) sending the sludge to land for agricultural application or (2)
producing struvite, which is recovered from the filtrates and/or centrates which have high
concentrations of NH4+-N and PO43--P. Since struvite has a greater per unit mass nutrient
content, struvite recovery appears to be a promising approach to nutrient recovery. Struvite
recovery potential case studies were therefore undertaken to examine how to increase the
efficiency of production and to investigate what other modifications could be made in the
upstream treatment process to increase the loading of the NH4+-N and PO43--P in the filtrates
and/or centrates.
The first case study was aimed at evaluating the impact of capturing greater amounts of
phosphorus in the sludge. Modelling of chemical dosing and enhanced biological phosphorus
removal (EBPR), to increase the loading of nutrient in the sludge dewatering centrates, was
undertaken. The case study suggests that EBPR is able to increase nutrient content in the sludge
centrate but that chemical dosing would not. This is because the chemically bonded phosphorus
is stable during sludge treatment. The majority of the extra phosphate captured will end up in
the dewatered digested sludge. In the EBPR process, the extra phosphorus captured in the
biological sludge is stored in the form of poly-phosphate, which will be released as phosphate
back to liquid phase under anaerobic conditions such as during sludge storage. Therefore, the
EBPR process is found to be able to boost the loading of the PO43--P in the centrate.
There are also three case studies conducted in Section 5.4 that aim to improve the energy
balance via modifying the wastewater treatment and sludge dewatering treatment. The
approaches are 1) improving the solid removal efficiency of the primary settlement process, 2)
reducing the sludge storage time, and 3) using energy generating secondary treatment. All
174
three were found to reduce the loading of the NH4+-N and PO43--P in the centrate. Approach 1)
and 3) reduced the nutrient captured in the secondary biological sludge and therefore, less
nutrient would be fed to the sludge treatment, and a greater percentage of nutrient would be
discharged with the final effluent. Approach 2) maintains the same amount of nutrient fed to
the sludge treatment. The reduction of sludge storage time, however, reduces the extent of
solubilization of the particulate nutrient and leads to a reduction of NH4+-N and PO43--P loading
in the centrate, but a greater nutrient reserve in the dewatered digested sludge.
175
Chapter 7 Business case study: pyrolysis of the biosolid for further energy
recovery
7.1 Introduction
Based on the mass and chemical energy balance constructed, the recovery opportunities for
nutrient and chemical energy are identified and examined in Section 5.4 and 6.3. This chapter
aims to choose the most promising approach and subsequently to develop a business case for
it.
The nutrient recovery from wastewater and sludge mainly focuses on struvite recovery from
the centrate from sludge dewatering. It can create income due to the fertilizer production and
may also reduce operational cost since chemical dosing for removal of the pipe scaling issue is
prevented (Khunjar et al., 2013, Ostara, 2010, de Vries et al., 2017, Molinos-Senante et al.,
2010). The economic feasibility of this has already been well demonstrated in different case
studies (Ostara, 2010, Kleemann et al., 2015).Thus, this study will not make further discussion
on it .
On the energy recovery side, Section 5.4 suggests that, improving the solid removal efficiency
of the primary settlement, pyrolysis of the dewatered digested sludge (called biosolid in this
chapter), and applying energy generating biological treatment can all potentially deliver a better
energy balance. But some approaches come with high uncertainty.
For primary settlement improvement, if it is delivered by metal chemical dosing, the stability
and digestibility of the sludge will be compromised (Diamantis et al., 2013).
For the energy generating biological treatment, no matter whether by AnMBR, or microbial
electrolysis cells and microbial fuel cells, none are yet proven to be completely energy self-
sustaining (Martin et al., 2011). In some cases, they even required a considerable amount of
energy input (Martin et al., 2011). Moreover, these technologies require a high influent COD
concentration but have limited ability in nutrient removal (Chernicharo, 2006). Therefore, they
are not a universal solution to all the treatment work, especially if the WWTP has a stringent
176
nutrient consent. They have not been previously implemented at full scale (Cotterill et al., 2017,
Williams et al., 2015, Shoener et al., 2014).
Among the proposed approaches for energy recovery, the biosolid pyrolysis has the least direct
impact to the current wastewater and sludge treatment process because it targets the end of the
sludge treatment and will not recirculate any waste stream back to the upstream treatment.
Moreover, because the agricultural land bank for anaerobic digestate is limited, the biosolid
has to compete with other organic source, i.e. food waste digestate (Bhogal et al., 2017). The
biosolid pyrolysis can reduce the volume of the digestate and hence reduce the land bank stress
(Paz-Ferreiro et al., 2018). Therefore, the biosolid pyrolysis is chosen as an example for
development into a business case study.
Brief of Base Case

The Base Case of this business case is the energy recovery system of WWTP C. WWTP C is a
district sludge treatment center having advanced anaerobic digestion (AAD). There is an onsite
combined heat and power (CHP) unit which is able to use the produced biomethane (CH4 in
the following content in this chapter) to produce electricity for the onsite wastewater treatment
and sludge treatment. But since 2014, the majority of the produced CH4 has been upgraded and
injected back to the gas grid. The drive for this act is that the renewable heat incentive (RHI)
claimed on the CH4 injected to the grid brings better cost benefit to the WWTP (personal
communication, L Wilkinson, Northumbrian Water, December 2016). Meanwhile, since CH4
is upgraded and sold, the CHP unit is mainly fed with imported natural gas. Also, part of the
electricity demand of the site is met by the import of electricity.
The aim of this business case study is to explore both the technical and economic feasibility of
the implementation of the biosolid pyrolysis to the Base Case energy recovery system.
7.2 Method
This study uses the payback period, net present value (NPV) and internal return rate (IRR) to
conduct the economic evaluation of the installation of the biosolid pyrolysis. This study
considers an economic feasible project should have a payback period shorter than the asset life,
an NPV that is greater than 0, and an IRR that is smaller than the nominal return rate.
177
7.2.1 The calculation of cash flows
The cash flow is the key for determining the payback period, the NPV and the IRR. The initial
cash flow is the capital expenditure (CAPEX) spent on the installation of the biosolid pyrolysis
system. The annual cash flow is determined by subtracting the operational expenditure from
the income. The operational expenditure (OPEX) considers the maintenance cost, labor and the
purchase of energy. The income considers the sale of the energy produced and the incentives
claimed. Both of the OPEX and income are related to the energy balance of the system.
The biosolid pyrolysis module is proposed to be installed at 2019 and is assumed to have an
asset life of 20 years, from 2019 to 2039. It worth noting that, the RHI for the CH4 injected to
the grid is valid for 20 years, from 2014 to 2034. This study assumes that after the RHI for the
CH4 injected to the grid expires after the first 15 years of asset life of biosolids pyrolysis, the
use of CH4 will change due to the economics of carrying on grid injection becoming weaker
(details will be explained in Section 7.3.3). This will result in a different energy balance from
the Base Case. Therefore, four energy balances are constructed for calculating the cash flow
based on whether the RHI for the CH4 injected to the grid is available or not, and they are
• EBGtG is the energy balance of the Base Case
• EBGtG+FP is the energy balance of the Base Case after the proposed fast pyrolysis is
installed
• EBCHP is the energy balance of the Base Case that has no CH4 grid injection and feeds
majority of CH4 to the CHP for electricity generation
• EBCHP+FP is the energy balance of the Base Case that has no CH4 grid injection, feeds
majority of CH4 to the CHP for electricity generation, and has a biosolid pyrolysis
Thus, four cash flow will then be constructed based on the above balance.
• CFGtG(I) is the cash flow calculated based on EBGtG and incentives being available
• CFGtG+FP(I) is the cash flow calculated based on EBGtG+FP and incentives being available
• CFCHP is the cash flow calculated based on EBCHP and no incentive is considered to be
available
• CFCHP+FP(I) is calculated based on EBCHP+FP and RHI incentive being only available to
be claimed on syngas produced by the biosolid pyrolysis
178
The real cash flow after the installation of biosolid pyrolysis for the first 15 years is
𝑪𝑭𝑮𝒕𝑮+𝑭𝑷(𝑰) − 𝑪𝑭𝑮𝒕𝑮(𝑰) . From the 16th year, after the current RHI for the CH4 injected to the
GtG is no longer available, the real cash flow becomes 𝑪𝑭𝑪𝑯𝑷+𝑭𝑷(𝑰) − 𝑪𝑭𝑪𝑯𝑷 .
7.2.2 The calculation of payback period, NPV and IRR
Payback period
The payback period is the expected time required to recover the capital expenditure (CAPEX).
Therefore, it is calculated by Equation 7-1.
Equation 7-1
𝐶𝐴𝑃𝐸𝑋
𝑃𝑎𝑦𝑏𝑎𝑐𝑘 𝑝𝑒𝑟𝑖𝑜𝑑 =
𝑅𝑒𝑎𝑙 𝑐𝑎𝑠ℎ 𝑓𝑙𝑜𝑤
A shorter payback period means the project has greater a liquidity (San Ong and Thum, 2013).
However, the payback period approach focuses solely on the capital recovery but ignore the
profit generate after the payback period (San Ong and Thum, 2013).
It is worth noting that the income is generated from the sale of energy or the incentive generated
from it, rather than from a fixed promised sum of cash given annually. The income, also the
OPEX, will be affected by the inflation. So is the cash flow. The payback period calculated in
this study has considered the time value of the money.
NPV
The core concept of NPV approach is to consider the time value of the money (Žižlavský,
2014). In most of the cases, it considers the same amount of money is worth less in the future
than today (Žižlavský, 2014) . This approach calculates the difference between the present
value of the inflowing cash and the present value of the outflowing cash of the investment to
be made (Žižlavský, 2014). If the NPV is greater than or equals to zero, the investment is
considered economic viable. If the NPV is negative, the investment is considered not economic
viable (Mack, 2012). The NPV is calculated via Equation 7-2 (Žižlavský, 2014).
179
Equation 7-2
𝑛
𝑁𝐶𝐹𝑡
NPV = ∑
(1 + 𝑟)𝑡
𝑡=0
Where 𝑵𝑪𝑭𝒕 in year t, 𝒓 is the real discount rate. In this study, the propose biosolid pyrolysis
system is assumed to have an asset life of 20 years. Therefore, the t is 20.
It is worth noting that the NPV is commonly used as a ranking tool to compare the cost benefits
of different proposed projects (Puchongkawarin et al., 2015). Because this study only proposes
one project, the NPV is only used as an economic indicator but not a ranking tool.
In the calculation of NPV, the real discount rate is an influential factor since it is used to
discount the cash inflow to today’s value. The real discount rate (𝑟𝑟𝑒𝑎𝑙 ) is decided by the interest
rate (𝑟𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡 ) and the inflation rate (𝑟𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛 ) via Equation 7-3 (Mack, 2012).
Equation 7-3
𝑟𝑛𝑜𝑚𝑖𝑛𝑎𝑙 𝑑𝑖𝑠𝑐𝑜𝑢𝑛𝑡 𝑟𝑎𝑡𝑒 − 𝑟𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛
𝑟𝑟𝑒𝑎𝑙 =
1 + 𝑟𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛
The interest rate is so called “the return on the investment” or “nominal cost of capital”. It is
named as nominal discount rate in this study. If the nominal discount rate is 20%, it means, the
minimum profit (made from the investment) required by the investor is 20% (Juhász, 2011).
The sponsor company commonly uses 12% as the interest rate for its investment project
(personal communication, Andrew Moore, Northumbrian Water, October 2017). Therefore,
this study uses 12% for the initial NPV calculation.
Inflation is a phenomenon that the prices of goods and services increases over time (Mack,
2012). According to Office of for National Statistics, the mean monthly inflation rate of the
UK during Jan 2017 to Dec 2018 is 2.43%. Therefore, this study takes 2.50% as the inflation
rate in all calculation.
IRR
The NPV only shows the increase in assets due to the investing during its life span but no
information about the yield of the profitability of the investment (Juhász, 2011). Meanwhile,
the IRR can (Juhász, 2011). IRR is the real discount rate when the NPV equals zero and then a
180
return of the investment is promised (Gallo, 2019). If the IRR is greater or equal to the nominal
discount rate, the investment is worthwhile (Gallo, 2019). In this study, the IRR is calculated
with the “irr” function of Microsoft Excel.
7.2.3 Sensitivity study on the uncertainty
As stated in Section 7.2.1, the cash flow calculation is based on the energy balance. The
energy balance construction in study is based on mass balance. Puig et al. (2008) reported
that poor data quality will results in erroneous mass balance. As stated in Section 5.4, the
mass balance could also change due to the modification of the process operation. Thus, there
is uncertainty embed in the mass balance. Besides, uncertainty also exist in the cost
estimation. Therefore, this study conducts sensitivity analysis to study how the uncertainty
could impact the economic evaluation.
This case study conducts sensitivity study on the following ten parameters,
(1) the amount of total solid in the biosolid
(2) the TS% of the biosolid
(3) the chemical energy content of the dried biosolid
(4) the electricity consumed on the fast pyrolysis of biosolid
(5) the efficiency of recovering syngas from biosolid. This will be called pyrolysis
efficiency in the following of content of this chapter
(6) the CAPEX
(7) the cost factor to estimate the maintenance cost,
(8) the additional labor
(9) the sale price of the surplus electricity
(10) the RHI claimed on the syngas
And the uncertainty of them are considered comes from

• For parameter (1), as stated in Section 5.4.3, the change of the upstream wastewater
and sludge process may affect the production of biosolid.
• For parameter (2), the TS% of biosolid is impacted by the dewatering process. For
parameter (3), the reported chemical energy content of the dried biosolid varies (Shizas
and Bagley, 2004, Schaum et al., 2016, Smyth et al., 2016).
181
• For parameter (4)-(8), they are estimated from literature value.
• For parameter (9), the sale price surplus electricity can vary depend on the deal with
the electricity supplier (personal communication, L Begg, OFGEM, January 2018).
• For parameter (10), this project did not propose an exact date for installation whilst the
RHI could change in the future.
A spider plot, which is a visualization aid, is used to present their impact to the pay back,
NPV and the IRR (El-Temtamy and Gendy, 2014). The approach is to give a ±20% variation
on one parameter once at a time, and then the newly obtained payback period, NPV and IRR
will be used to plot against the ratio of the changed value to the original value of that
parameter (El-Temtamy and Gendy, 2014, Lipu and Jamal, 2013). If a line of the
corresponding parameter has a steep slope, then the parameter is considered to have a big
influence on the predicted pay back, NPV and IRR (Lipu and Jamal, 2013, El-Temtamy and
Gendy, 2014).
The found most influential parameter(s) will be chosen for being further studied the payback
period, NPV, and IRR under a pessimistic scenario and an optimistic scenario (Rysanek and
Choudhary, 2013). In this study, the pessimistic scenario is that parameter(s) is(are) at the
worst condition, and the optimistic condition is that all three parameter(s) is(are) at the best
condition. The scenario study is for deeply understand the impact from the uncertainty to the
economic evaluation.
7.3 Result and Discussion
7.3.1 Energy balance
The data shown below in this Sub-section 7.3.1 is the daily average collected from sludge
treatment of WWTP C from 01/03/2016 to 31/08/2016. The data is collected from
Northumbrian Water (personal communication, S Coverdale, Northumbrian Water, October
2017 and personal communication, L Wilkinson, Northumbrian Water, October 2018).
182
7.3.1.1 Energy balance of the energy recovery system (EBGtG) of the Base Case
In the Base Case, the energy recovery system of the WWTP C consist of three major
components, the boiler, the CHP, and the Gas to Grid (GtG) upgrade plant. The input energy
containing materials are the indigenously produced biomethane (CH4), the natural gas
purchased from the grid, and the purchased propane. Also, a certain amount of electricity is
purchased for meeting the demand of the sludge treatment and wastewater treatment. There are
four outputs of the system, the steam from the boiler, the low-grade heat from the CHP, the
electricity produced from the CHP, and the upgraded gas from the GtG upgrade plant. The
produced steam will be used for the thermal hydrolysis process (THP) process of the AAD.
The electricity will be consumed onsite for sludge and wastewater treatment. The low-grade
heat is usually wasted. The produced upgrade gas will be injected to the gas grid.
In detail, The CH4 produced from the AD was measured as 245.5 MWh. Since the CH4 flared
was measured as 20.7 MWh, only 224.8 MWh CH4 enters into the system, along with 257.1
MWh natural gas and 23.7 MWh propane.
27.6 MWh of CH4 and 212.8 MWh of natural gas are fed to the CHP. The electricity production
is measured as 83.2 MWh. The heat to electricity conversion is 34.6%. It is worth noting that,
9.6 out of 83.2 MWh electricity is generated from the 27.6 MWh CH4. Based on the literature,
the conversion ratio of both high-grade heat and low-grade heat is assumed as 20% (Smyth et
al., 2016, Bowen et al., 2010). Therefore, 48.1 MWh of high-grade heat and 48.1 MWh of low-
grade heat are then produced.
48.1 MWh of high-grade heat made up with 44.3 MWh of natural gas are fed to the boiler. The
heat production is measured as 77.3 MWh of steam. The energy recovery ratio is 83.7% which
agrees with the reported value approximately 85% (Vanwortswinkel and Nijs, 2010).
The rest 197.2 CH4 is fed to the GtG upgrade plant along with 23.7 MWh propane. 220.9 MWh
upgrade gas is injected into the grid.
In the end, the system produces 77.3 MWh steam which meets the demand of the THP, 48.1
MWh low-grade heat, 220.9 MWh upgrade gas, and 83.2 MWh electricity which is still 18.8
183
MWh short to fulfil the daily electricity demand measured as 102.0 MWh. The energy balance
of the Base Case is shown in Figure 7-1.
The cash flow related energy flow on a daily basis is summarized as

• Purchase: 257.1 MWh natural gas, 23.7 MWh propane, and 18.8 MWh electricity
• Sale: 220.9 MWh upgrade gas
• Incentives received: 197.2 MWh CH4 for the RHI claim, and 9.6 MWh electricity for
the Renewable Obligation Certificate (ROC) trade (personal communication, Steve
Coverdale, Northumbrian Water, October 2017).
Figure 7-1 The energy balance (EBGtG) of the Base Case
7.3.1.2 Energy balance of the Base Case energy recovery system with the addition of biosolid
pyrolysis (EBGtG+FP)
The design of sludge fast pyrolysis has three units; the sludge drying, which utilizes the low-
grade heat to dry the sludge to reach 90% DS (Mills et al., 2014), the fast pyrolysis to convert
biosolid into syngas, and the CHP for combusting the syngas produced for electricity and heat
production. Apart from the original three inputs, CH4, natural gas, and the propane, the energy
containing biosolid is fed to the system. The installation of the pyrolysis requires no
modification of the Base Case energy system and its operation. The output remains the same
for, steam, low-grade heat, upgrade gas, and electricity.
184
The energy balance is presented on a daily basis. 41 tonnes dried solid (DS) biosolid with 177.7
MWh of chemical energy is fed to pyrolysis. Reportedly, 0.77-1.09 MWh heat is required to
evaporate 1 ton of water (Mills et al., 2014). The heat supplied can be from the low-grade heat
from the CHP (Mills et al., 2014). This study assumes the ratio is 0.9 MWh heat per ton water
evaporated. In the sampling campaign conducted in the mass and chemical energy balance
investigation, 5 times the TS% of the biosolid were measured all were 28.5%. When the TS of
the biosolid is increased from 28.5% to 90.0%, 41÷28.5%-41÷90.0%= 98.3 tonne water is
removed and hence 0.9×98.3=88.5 MWh of low-grade heat is required. Whilst 0.055 MWh
electricity is required to process each tonne DS content (Mills et al., 2014). Therefore, 88.5
MWh/day of low-grade heat and 2.3 MWh/day electricity is required for drying the sludge.
Approximately 60%-90% of the chemical energy within the biosolid can be recovered as
syngas and fed to the CHP, the rest is lost or remains in the bio-char (Mills et al., 2014, Cao
and Pawłowski, 2012). This study assumes the conversion ratio is 80%. Therefore, the thermal
energy within the syngas is 177.7×80.0% = 142.1 MWh. The pyrolysis does require electricity
input for the heating. According to Mills et al. (2014), this study assumes the ratio is 0.4 MWh
electricity per ton DS. Therefore 16.4 MWh electricity is required.
The produced syngas is then fed to the CHP. Assuming using the CHP for the syngas recovers
38% of the 142.1 MWh of influent energy as electricity, then electricity production is 54.0
MWh/day (Mills et al., 2014). Meanwhile, the conversion ratio of both high-grade heat and
low-grade heat are assumed to be 20% (Smyth et al., 2016, Bowen et al., 2010). Therefore,
28.4 MWh of high-grade heat and 28.4 MWh of low-grade heat are also produced.
The sludge pyrolysis produces 54.0 MWh electricity from the daily 41.0 tonnes DS biosolid.
After offsetting the 2.3+16.4=18.7 MWh parasitic load of electricity consumed in sludge drying
and pyrolysis unit, the net electricity output of the biosolid pyrolysis module is 35.3 MWh. The
system also produces 56.8 MWh heat, but it is not able meet the demand of 88.5 MWh for
sludge drying. The heat deficit of the biosolid pyrolysis module is 31.6 MWh per day.
With integration to the Base Case energy recovery unit, the unused 48.1 MWh low-grade heat
from the original system can be used to fulfil the 31.6 MWh heat deficit of the gasification
185
system, with still 16.5 MWh surplus. Meanwhile, the total electricity output is increase to 118.5
MWh. After offsetting the 102.0 demand, there is still 16.5 MWh net electricity output which
can be sold to the electricity grid. The production of the steam and upgrades gas is remained at
77.3 MWh and 220.9 MWh upgrade gas, respectively.
The cash flow related energy flow on a daily basis becomes:

• Purchase: 257.1 MWh natural gas and 23.7 MWh propane,
• Sale: 220.9 MWh upgrade gas and 16.5 MWh electricity
• Incentives received: 197.2 MWh CH4 for the original RHI claim, and 9.6 MWh
electricity for the Renewable Obligation Certificate (ROC) trade (personal
communication, Steve Coverdale, Northumbrian Water, October 2017), and142.1 MWh
of syngas for the new RHI claim.
The energy balance of the Base Case with biosolid pyrolysis is shown in Figure 7-2.
Figure 7-2 The energy balance (EBGtG+FP) of the Base Case system with biosolid pyrolysis
186
7.3.1.3 Energy balance of the Base Case energy recovery system with no GtG (EBCHP)
Since the RHI will end within the project life time of pyrolysis unit, modification of the Base
Case system is likely to occur. This study assumes the change will be diverting the CH4 back
to the CHP for electricity production. The reason is that, the buying price of the natural gas
which current used for electricity production is £13/MWh, but the sale price of the upgrade gas
is £10/MWh while the extra cost for the purchase of propane is involved (personal
communication, Steve Coverdale, Northumbrian Water, October 2017).
However, in the original EBGtG, the natural gas usage is 212.8 MWh and the CH4 fed to the
GtG upgrade of only 197.2 MWh. There is a shortage of 15.6 MWh chemical energy for the
CHP. Based on the energy conversion ratio of the current CHP is 34.6%, 15.6 MWh heat energy
is equivalent to 5.4 MWh electricity. Either 15.6 MWh natural gas or 5.4 MWh electricity has
to be brought from the grid. The buying price of natural gas is only £13/MWh but of electricity
is £114/MWh. Thus, in a daily basis, the import of natural gas costs £15.6×13=£202.8, but the
import of electricity costs £114×5.6=£615.6. It is therefore more economic to buy 15.6 MWh
natural gas to produce electricity on site. Therefore, this study assumes 15.6 MWh natural gas
will be brought from the grid per day.
After the route of GtG is abandoned at the end of the RHI incentive to the CH4 to grid, the
input to system are only the 59.9 MWh natural gas and 224.8 MWh CH4. The output of the
system becomes 77.3 MWh steam which still meets the demand of the THP, 48.1 MWh low-
grade heat, and 83.2 MWh electricity which is still 18.8 MWh short of fulfilling the daily
electricity demand measured as 102.0 MWh.
The cash related energy flow on a daily basis is summarized as

• Purchase: 59.9 MWh natural gas and 18.8 MWh electricity
The energy balance of the Base Case system with no GtG is shown in Figure 7-3.
187
Figure 7-3 The energy balance (EBCHP) of the system has no CH4 grid injection
7.3.1.4 Energy balance of the Base Case energy recovery system with biosolid pyrolysis but
with no GtG (EBCHP+FP)
The end of the GtG route does not affect low grade heat production and hence will not affect
the performance of the energy balance of the pyrolysis module as stated in Section 7.3.1.4. The
pyrolysis module still has a net output of the electricity of 35.3 MWh per day and suffers a heat
deficit is 31.6 MWh per day. Once the module is added on to the system with no GtG, the heat
deficit will be rectified, and the net electricity output will offset the 18.8 MWh electricity
import. Although there is still surplus electricity after the offset, this electricity is preferable to
be sold to the grid rather than supplementing the gap due to the CH4 shortage stated in Section
7.3.1.3. This is because 2.9 MWh of natural gas is needed for generating 1 MWh electricity
(based on the heat to electricity conversion ratio is 34.6%) and the cost is only £13×2.9= 37.7,
whilst the sale price of 1 MWh of electricity is £48.5. Therefore, it is more economic to carry
on importing the natural gas for electricity production.
After the route of GtG is abandoned but a pyrolysis module is added in, the input to system are
still 59.9 MWh natural gas and 224.8 MWh CH4. The output of the system becomes 77.3 MW
of steam, 16.5 MWh low-grade heat, and 16.5 MWh electricity.
The cash related energy flow on a daily basis is summarised as

• Purchase: 59.9 MWh natural gas
• Sale: 16.5 MWh electricity
188
• Incentives received: 142.1 MWh of syngas for the new RHI claim.
The energy balance of the Base Case system with no GtG but has biosolid pyrolysis is shown
in Figure 7-4.
Figure 7-4 The energy balance (EBCHP+FP) of the system has no CH4 grid injection but with biosolid pyrolysis
7.3.2 Cash flow
This subsection will first present the estimation and calculation approach to the CAPEX, OPEX
and income, per equipment or, per unit of energy as in Section 7.3.2.1, 7.3.2.2, and 7.3.2.3,
respectively. Then the real cash flow will be shown in Section 7.3.2.4.
7.3.2.1 Capital Expenditure

This study assumes there is no CAPEX involved in the Base Case with no pyrolysis module.
The CAPEX of the pyrolysis module is spent on the installation of the sludge drying plant, the
gasifier, and the CHP for Syn-Gas. The CAPEX of the unit (𝐶𝑎𝑝𝑒𝑥𝑝 ) is estimated via Equation
7-4 (Jones et al., 2013).
Equation 7-4
𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑝 𝑛
𝐶𝑎𝑝𝑒𝑥𝑝 = 𝐶𝑎𝑝𝑒𝑥𝑟 × ( )
𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑟
Where 𝑪𝒂𝒑𝒆𝒙𝒓 is capital cost of the similar process unit reported in literature, 𝑪𝒂𝒑𝒂𝒄𝒊𝒕𝒚𝒑 is
the capacity of the proposed unit, 𝑪𝒂𝒑𝒂𝒄𝒊𝒕𝒚𝒓 is the capacity of the similar process unit reported
189
in literature, and n is the scale factor, commonly, 0.6-0.7 (Jones et al., 2013). This study takes
the n as 0.65.
As stated in Section 7.3.1.2, 98.3 tonnes water is expected to be evaporated, 41 tonnes of TDS
are expected to be treated, and 54 MWh electricity (equivalent to a power rating of
54,000kWh÷24h=2250 kW) is expected to be produced. The estimated capital cost of the three
process units are listed in Table 7-1.
Table 7-1 The estimated CAPEX of the components of the biosolid pyrolysis
Reference 𝑪𝒂𝒑𝒆𝒙𝒓 , £ 𝑪𝒂𝒑𝒂𝒄𝒊𝒕𝒚𝒓 𝑪𝒂𝒑𝒂𝒄𝒊𝒕𝒚𝒑 𝑪𝒂𝒑𝒆𝒙𝒑 , £

Sludge Drying
Mills et al. (2014) 3,676,922 77 tH2O/d* 98.3 tH2O/d 4,309,600
Fast Pyrolysis
Mills et al. (2014) 8,120,331 60 tds/d* 41 tds/d 6,339,937
Jones et al. (2013) 123,120,000 2000 tds/d 41 tds/d 9,839,400
CHP
Mills et al. (2014) 7,504,802 4000 kW 2250 kW 5,163,862
BEIS (2017a) 5,300,000 7700 kW 2250 kW 2,382,491
BEIS (2017b) 2,700,000 1400 kW 2250 kW 3,675,826
BEIS (2017c) 7,300,000 4000 kW 2250 kW 5,022,943
*: tH2O/d stands for ton water evaporated per day, and tds/d stands for ton dried solid per day.
The estimated CAPEX of the sludge drying plant is £ 4,309,600. The estimated capital cost of
the fast pyrolysis and the CHP are taken from the mean value of the multiple estimated values,
and they are £8,089,668 and £4,060,281, respectively. Therefore, the total CAPEX of the
biosolid pyrolysis system is estimated as £16,460,549 but could vary from £13,032,028 to
£19,312,862
7.3.2.2 Operational expenditure

In this study, the operational expenditure (OPEX) includes maintenance, labour cost, and the
material, namely the purchase of natural gas, propane and electricity.
Maintenance
The typical maintenance costs can be estimated as 1-3% of the capital cost. This study assumes
the ratio is 2% since the energy consumption and the labor cost is calculated separately. The
190
investment in the gas to grid facility at WWTP C is reported to have amounted to £8,000,000
(WWT online, 2015). The annual maintenance cost is estimated as £160,000. For the original
CHP, its maintenance is calculated based on a ratio of £6 per MWh electricity produced.
Therefore, the maintenance cost per annum is 6×83.2×365= £182,208 (DECC, 2008).
For the biosolid pyrolysis module, based on the CAPEX estimated in Section 7.3.2.1, the
maintenance cost of the drying plant is £86,190, of the fast pyrolysis is £161,793, and of the
syngas CHP is £81,215, respectively.
Labor
The annual salary of a wastewater operator is from £14,500 to 32,000 (OECD, 2017). The labor
cost of a GtG process is suggested as £31,000 (DECC, 2014). Thus, their wages are assumed
to be £30,000 per person per annum.
Currently, there are 18 operators working for this Base Case energy system that processes 85
tonnes DS per day (based on the average from 10/03/2016 to 31/08/2016) (personal
communication, John Robinson, Northumbrian Water, October 2017). This study assumes the
labor needed is proportional to the amount of dried solid processed. Hence, the 9 additional
operators could be required for the sludge gasification system that treats 41 tonnes DS.
Considering the operators could be allocated from other operation teams, this study assumes 5
instead of 9 operators will be added to the operation team of the energy recovery system
(personal communication, A Moore, Northumbrian Water, January 2019).
Material
The buying price of natural gas is £ 13.0 per MWh and the buying price of propane is £100.0
per MWh (personal communication, S Coverdale, Northumbrian Water, October 2017).
According to statistic “Prices of fuels purchased by non-domestic consumers in the UK”
published by Department for Business, Energy & Industrial Strategy, the electricity sale price
to a medium size (2,000-19,999 MWh) consumer is 114.4 per MWh. Since the electricity
import of WWTP C is 18.8×365=6,862 MWh per annum, it should be categorized as medium
size consumer. Therefore, the buying price of electricity is assumed as £114.4 per MWh.
191
7.3.2.3 Income
The income includes the sale of exported gas and electricity to the gas grid, and the incentives
received.
The upgrade gas injected to grid is sold at £ 10.0 per MWh (personal communication, S
Coverdale, Northumbrian Water, October 2017). Moreover, the sale of electricity to the grid is
estimated as price £52.4 per MWh, the same as the export price of the Feed-In-Tariff (OFGEM,
2019a).
The Renewable Heat Incentive (RHI) is available to claim on the CH4 injected to the grid. The
unit price for WWTP C is £ 76.2 per MWh (personal communication, S Coverdale,
Northumbrian Water, October 2017). It is also available to claim on the syngas produced by
the biosolid pyrolysis and injected to the CHP (OFGEM, 2018). Therefore, this study assumes
the pyrolysis and CHP is accredited for the RHI, and £11.6 per MWh is gained for the syngas
fed to the CHP (OFGEM, 2019b).
A Renewable Obligation Certificate (ROC) can be granted for the electricity produced from
the CH4 in the exiting CHP. The trade price of ROC is £ 42 per MWh electricity (personal
communication, S Coverdale, Northumbrian Water, October 2017).
The biosolid produced at WWTP C is currently sold to the farmers. Considering the biochar,
the resulting product of the pyrolysis, may have the similar fertility with biosolid, this study
assumes the income generated from the sale of biosolid and from the sale of the biochar are the
same, and hence are not considered in the cash flow.
192
7.3.2.4 Real cash flow
The annual cash flows of different energy balances are shown in Table 7-2.
Table 7-2 The annual cash flow of different energy balance and under different scenarios
CFGtG(I) CFGtG+FP(I) CFCHP CFCHP+FP(I)
Operational Expenditure
Purchase of Electricity £785,013 £0 £785,013 £0
Purchase of Natural Gas £1,219,940 £1,219,940 £284,226 £284,226
Purchase of Propane £865,050 £865,050 £0 £0
Labor cost £540,000 £540,000 £540,000 £540,000
Maintenance (CHP) £182,208 £182,208 £182,208 £182,208
Maintenance (GtG) £160,000 £160,000 £0 £0
Labor cost (Fast Pyrolysis) £0 £150,000 £0 £150,000
Maintenance (Sludge Drying Plant) £0 £86,192 £0 £86,192
Maintenance (Fast Pyrolysis) £0 £161,793 £0 £161,793
Maintenance (Fast Pyrolysis CHP) £0 £81,226 £0 £81,226
Total Operational Expenditure: £3,752,210 £3,446,408 £1,791,446 £1,485,644
Income
Sale of Upgraded Gas £806,285 £806,285 £0 £0
RHI claimed on the CH4 of Upgraded Gas £5,484,724 £5,484,724 £0 £0
ROC claimed on Electricity produced by
£146,433 £146,433 £0 £0
Biomethane
Sale of Electricity £0 £316,644 £0 £316,644
RHI claimed on the Syngas to CHP £0 £601,793 £0 £601,793
Total Income: £6,437,442 £7,355,878 £0 £918,436
Cash flow： £2,685,232 £3,909,470 -£1,791,446 -£567,208
No matter whether the GtG exists or not, the installation of the biosolid pyrolysis is able to
reduce the OPEX by £305,802 per annum which is attributed to the scrapping of the purchase
of electricity. Also, it can bring in an extra income of £918,436 from the sale of surplus
electricity and the RHI claimed in the syngas. Therefore, the real cash flow of the first 15 years
and last 5 years remains the same as £1,224,238.
It is worth noting that, apart from CFCHP, 65.5-87.5% of the income comes from the incentives
(Figure 7-5). As shown in Table 7-2, if the incentives are scrapped, both CFGtG(I) and CFGtG+FP(I)
will become negative because the income from the sale of the upgrade gas and electricity is far
too low to cover the OPEX. The incentives are vital to the cash flow.
193
CF(GtG(I))
CF(GtG+FP(I))
CF(CHP+FP(I))
0% 20% 40% 60% 80% 100%
RHI claimed on the CH4 of Upgraded Gas

RHI claimed on the Syngas to CHP
ROC claimed on Electricity produced by Biomethane
Sale of Upgraded Gas
Figure 7-5 The percentage breakdown of income of different cash flow
7.3.3 The economic evaluation and the sensitivity study on the uncertainties
7.3.3.1 The economic evaluation

Base on the real cash flow, the payback period, the NPV and the IRR of the installation of the
biosolid pyrolysis system are calculated as 13.4 years, -£5,495,459, and 4.12%, respectively.
Although the payback period suggests the investment could be recovered during the asset life,
the negative NPV and an IRR that smaller than the nominal discount rate (12%) both suggesting
that the installation of biosolid pyrolysis is not currently economically feasible.
7.3.3.2 The sensitivity study on the uncertainties

Ten parameters with uncertainties are studied for their impact to the economic evaluation.
Figure 7-6 to 7-8 shows the spider plot of the sensitivity study.
20
Payback, year
15
10
0
80% 90% 100% 110% 120%
Ratio to original value
Solid Loading, ton/d TS% of Biosolid
Chemical energy of dried biosolid Electricity used for per ton of solid
Pyrolysis Efficiency Capex
Cost factor for Maintenance Additional Labour
Sale Price of surplus electricity RHI claimed on Syngas
Figure 7-6 The spider chart of payback period
194
0
80% 90% 100% 110% 120%
-2000000
-4000000
NPV, £
-6000000
-8000000
-10000000
Ratio of original value
Figure 7-7 The spider chart of the NPV
8.0%
7.0%
6.0%
5.0%
IRR
4.0%
3.0%
2.0%
1.0%
0.0%
80.0% 90.0% 100.0% 110.0% 120.0%
Ratio to original value

Figure 7-8 The spider chart of the IRR
The increase of TS% of biosolid, chemical energy of dried biosolid, pyrolysis efficiency, sale
price of surplus electricity and RHI claimed on Syngas will leads to a higher NPV, a greater
IRR and a shorter payback period. The impact caused by a 20% increase of these parameters
(once at a time) to NPV, IRR and payback period are listed in Table 7-3.
Meanwhile, the decrease of electricity used for per ton of solid (in the pyrolysis), CAPEX, cost
factor of maintenance, additional labour, sale price of surplus electricity and RHI claimed on
Syngas will also leads to a higher NPV, a greater IRR and a shorter payback period. The impact
195
caused by a 20% decrease of these parameters (once at a time) to NPV, IRR and payback period
are also listed in Table 7-3.
Interestingly, the increase of solid loading results in a lower NPV but a higher IRR and a shorter
payback period. In the cash flow calculation, the CAPEX is estimated base of solid loading,
water loading in the sludge, and the projected electricity production. The maintenance cost is
then estimated from the CAPEX. The increase of the solid loading will increase the CAPEX
and subsequently increase the maintenance cost. However, it also leads to a greater production
of syngas and electricity. The income from the RHI claimed on the syngas and sale of surplus
electricity is therefore elevated. In result, if the solid loading is increased by 20%, the CAPEX
will be increased from £16,460,549 to £18,351,566 and the cash flow will also be increased to
£1,224,238 to £1,438,419. The increase of both CAPEX and cash flow explains why a greater
solid loading will lead to a lower NPV but a higher IRR and a shorter payback period. However,
a 20% increase of the solid loading only leads to a 5% worse NPV, 12% better IRR and 4%
shorter payback period.
Table 7-3 The impact to the NPV, IRR and Payback period if a ±20% change is given to the parameters tested
for sensitivity
Improvement Improvement Payback Improvement

NPV IRR
on NPV* on IRR* Period on Payback*
TS% of Biosolid** -4,651,032 15% 4.77% 16% 12.7 5%
Chemical energy of
-3,169,498 42% 6.51% 58% 11.0 18%
dried biosolid**
Pyrolysis Efficiency** -3,169,498 42% 6.51% 58% 11.0 18%
Sale Price of surplus
-4,928,245 10% 4.70% 14% 12.8 5%
electricity**
RHI claimed on
-4,417,448 20% 5.21% 26% 12.2 9%
Syngas**
Electricity used for per
-4,933,578 10% 4.70% 14% 12.8 5%
ton of solid***
Capex*** -1,613,623 71% 7.48% 82% 10.2 24%
Cost factor for
-4,905,732 11% 4.72% 15% 12.8 5%
Maintenance***
Additional Labor*** -5,226,759 5% 4.40% 7% 13.1 2%
𝑈𝑝𝑑𝑎𝑡𝑒𝑑 𝑁𝑃𝑉,𝐼𝑅𝑅,𝑜𝑟 𝑃𝑎𝑦𝑏𝑎𝑐𝑘−𝑂𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑁𝑃𝑉,𝐼𝑅𝑅,𝑜𝑟 𝑃𝑎𝑦𝑏𝑎𝑐𝑘
*: The improvement is calculated by | | × 100%
𝑂𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑁𝑃𝑉,𝐼𝑅𝑅,𝑜𝑟 𝑃𝑎𝑦𝑏𝑎𝑐𝑘
**: A 20% increase is given to each of these parameters
***: A 20% decrease is given to each of these parameters
196
For all three economic feasibility indicators, the CAPEX, the chemical energy content of dried
biosolid and the pyrolysis efficiency are the most influential parameters, since their lines have
the steepest slope as shown in Figure 7-6 to 7-8. This is because the calculation of the payback
period, NPV and IRR are affected by three elements, the CAPEX, the OPEX and the income,
and these three parameters are able to simultaneously influence two elements or have a double
effect on one element. For the CAPEX, its change does not only affect the amount of
investment needed to be paid but also affects the cash flow since the maintenance cost of the
newly installed facility is estimated from the capital. For the chemical energy content of dried
biosolid and the pyrolysis efficiency, they affect the production of the syngas, and hence
influences the RHI that could be claimed from the syngas fed to the CHP. Moreover, the
increased production of the syngas will also positively affect the production of electricity and
hence the income of the sale of electricity. But the others can only affect one element at a time.
The TS% of the biosolid only affects the CAPEX estimated for the sludge drying plant. The
sale price of surplus electricity, RHI claimed on syngas and electricity used for pyrolysis affects
the income only, whilst the cost factor for maintenance and additional labour just impacts the
OPEX.
The pessimistic scenario and optimistic scenario study are conducted on the three most
influential parameters, the CAPEX, the pyrolysis efficiency and the chemical energy content
of dried biosolid are found as the most influential factors in the economic evaluation.
In literatures, the reported pyrolysis efficiency is between 60% to 90% (Cao and Pawłowski,
2012, Mills et al., 2014). For the chemical energy content of dried biosolid, the reported values
range for 11.1 to 15.7 kJ/g (Shizas and Bagley, 2004, Mills et al., 2014, Schaum et al., 2016).
In the CAPEX estimation shown in Section 7.3.2.1, the CAPEX of the CHP is estimated from
electricity production which is influenced by the pyrolysis efficiency and the chemical energy
content of dried biosolid. Different pyrolysis efficiency and the chemical energy content of
dried biosolid will results in different estimated CAPEX. In this study, the CAPEX in the
pessimistic scenario is its highest estimation based on the worst pyrolysis efficiency and the
chemical energy content of dried biosolid. The CAPEX in the optimistic scenario is in the
opposite. Therefore,
197
• In the pessimistic scenario, the CAPEX is assumed as £17,582,155, the pyrolysis is
assumed as 60%, and the chemical energy content of dried biosolid is assumed as 11.1
kJ/g. Based on the pessimistic scenario, the payback period, NPV and IRR are
calculated 40.0 years, -£13,646,391, and -5.86%, respectively.
• In the optimistic scenario, the CAPEX is assumed as £13,232,298, the pyrolysis is

assumed as 90%, and the chemical energy content of dried biosolid is assumed as 15.7
kJ/g. Based on the optimistic scenario, the payback period is calculated 8.8 years, NPV
becomes a positive value of £ 246,696, but the IRR is 9.53% which is still smaller that
the nominal discount rate of 12% , respectively.
Although the economic feasibility is not proven in both pessimistic and optimistic scenarios,
there can be a substantial difference in the predicted payback period, NPV and IRR in this
proposed project. For managing this huge uncertainty, the Monte Carlo analysis can be used
(Korytárová and Pospíšilová, 2015). The analysis first give numbers randomly chosen (from a
reasonable range) to the influential parameter to obtain large amount of the predicted economic
indicators, and then work out the probabilistic distribution of the economic indicators (as
shown in Figure 7-9) (Korytárová and Pospíšilová, 2015). The result can then demonstrate the
economic feasibility under consideration of risks. However, since the economic feasibility is
not proven in the optimistic scenario, the Monte Carlo analysis is not performed in this study.
Figure 7-9 The probabilistic distribution of NPV obtained from Monte Carlo Analysis (Korytárová and
Pospíšilová, 2015)
198
7.3.3.3 Future opportunities
This section is to discuss at what circumstance the economics of this proposed project will
become feasible. The discussion is made on the optimistic scenario discussed in Section 7.3.3.2.
Each of the three most influential parameters will be adjusted once at a time till a payback
period within 20 years, a positive NPV and an IRR greater than 12% are achieved
simultaneously.
For the CAPEX, the economic feasibility will be achieved if it is reduced to lower than
£11,499,245 which is £1,739,052 less than the optimistic estimation of £13,232,298.
Potentially this may happen because the future technology improvement can reduce the
CAPEX of the facility, for example, the capital cost of wind turbine electricity generators at
present has fallen by approximately 20% comparing to in 2008 due to the technology
improvement (IRENA, 2016).
For the pyrolysis efficiency, even it is raised from the optimistic estimation of 90% to 100%,
the IRR obtained is still 11.22%. It is not able to achieve the economic feasibility. Pyrolysis
efficiency is sensitive to the economic evaluation. But in this study, it has been assumed at high
level (80%) and is with less room to improve. It cannot single-handedly elevate the economic
of the proposed project to the desired level.
For the chemical energy content of the biosolid, if it is raised to 18.3 kJ/g, the economic
feasibility will be achieved. However, the chemical energy content of the dried digested sludge
is reported 11.1 to 15.7 kJ/g and is not likely to be as high as 18.3 kJ/g (Schaum et al., 2016,
Smyth et al., 2016, Shizas and Bagley, 2004). But such a high kJ/g value can be found on the
non-digested sludge as shown in Table 4-6. This implies that non-digested sludge may be a
more suitable feed of the pyrolysis.
Besides, although incentive are not classified as the most influential parameter in this study, it
plays an important role in determining the economics of high capital related projects (Mills et
al., 2014). The best example is the current operating GtG in the Base Case whose economics
substantially relies on the RHI claimed on the CH4 injected to the grid (personnel
communication, L Wilkinson, Northumbrian Water, December 2016). Also, as stated in the
Section 7.3.2, the 65.5% projected income of the proposed project is from the incentives
199
claimed on the syngas. If future market favors generating energy from biosolid or solid waste
via pyrolysis, the better economics could be achieved. In particular, the change of incentives
tariff is sometimes dramatic. According to the historical data publish by OFGEM, for the
“Large biogas combustion” applied in this project, the highest ever announced tariff was £24.3
per MWh for the CH4 combusted and the lowest was £8.9 per MWh (OFGEM, 2019b).
Applying the £24.3 per MWh incentive to the optimistic scenario, the projected payback period
becomes 5.5 years, the NPV is improved to £6,035,248, and the IRR is 15.65%. The biosolid
pyrolysis may have a future if incentives will become better.
7.3.4 Environmental Benefit
The greatest environmental benefit brought about by the biosolid pyrolysis plant is the
improved energy balance of the Base Case. The energy balance EBCHP (stated in Section 7.3.1.3)
and EBCHP+FP (stated in Section 7.3.1.3) that are shown without GtG are used for further
demonstration. These two comparative balances are in a unified energy type, electricity. All the
input or outputs of natural gas will be converted into electricity at the conversion ratio of 38%.
The energy figures shown in the following content is given on a daily basis.
In EBCHP, the system has to import 59.9 MWh natural gas and 18.8 MWh electricity but
generates no output energy. Therefore, the Base Case is a net energy importer and the electricity
balance is -59.9×38%+(-18.8)= -41.6 MWh. For EBCHP+FP, the system with biosolid pyrolysis
still imports 59.9 MWh natural but no electricity. Moreover, the system can even export 16.6
MWh electricity back to the grid. Although the system is still a net energy importer, the
electricity balance is improved to -59.9×38%+16.6= -6.2 MWh. The total electricity saving is
35.4 MWh electricity. In 2017, the CO2 emitted from per MWh electricity supplied is 0.2 ton
(BEIS, 2018). Hence, the annual CO2 emission reduction will be 35.4×0.2×365=2584.2 tonnes.
Furthermore, the biochar comes with smaller volume than the biosolid (Paz-Ferreiro et al.,
2018). Implementation of biosolid pyrolysis can also reduce the energy consumption of the
biosolid haulage.
Moreover, currently biosolid is mainly sent for land application due to its nutrient content. But
it also has some potential environmental issues, such as the ammonia gas emission and the
concern of possible pathogen regrowth on land, etc.(Paz-Ferreiro et al., 2018, Higgins and
200
Murthy, 2006). The biochar, the resulting product of pyrolysis, still has a comparable fertility
to biosolid but is free from these issues since pyrolysis is a high temperature treatment which
can evaporate the free ammonia and kill pathogen (Paz-Ferreiro et al., 2018, Wang et al., 2012).
Pyrolysis can unlock the energy reserve of the biosolid, prevent any adverse effect of its land
application, but still retain its advantageous fertility.
7.4 Chapter Summary
Since the biosolid pyrolysis potentially has the least impact to the upstream wastewater and
sludge treatment process and could potentially help to reduce the landbank stress, this chapter
investigated the technical and economic feasibility of its implementation on to the existing
energy recovery system of WWTP C.
This chapter first studied the energy balance of the Base Case system and the system with the
addition of the biosolid pyrolysis. The economic evaluation considering the payback period,
NPV and IRR were then made according to the energy balance. The result suggested that the
installation of biosolid pyrolysis is not currently economically feasible.
In the sensitivity study of the uncertainty, CAPEX, pyrolysis efficiency and the chemical
energy content of the dried biosolid were found the most influential parameters to the economic
evaluation. Even the three were given an optimistic estimation, but the economic feasibility is
still not achieved. However, if the CAPEX can be further reduced or better incentives will be
granted in the future, the installation of biosolid pyrolysis could become economically feasible.
Although biosolid pyrolysis is not currently economically feasible, it can still bring in
environmental benefits of 1) a better energy balance of the system, 2) a reduced energy
consumption on sludge haulage, and 3) the prevention of the ammonia gas emission and the
possible pathogen regrowth from the biosolid.
201
Chapter 8 Conclusion and Recommendation
8.1 Conclusion
The overall aim of this project is to reliably determine the chemical energy and mass balances
of wastewater treatment plants in the UK to inform possible options for recovery of organic
carbon (for energy), nitrogen and phosphorus.
To fulfil this overall aim, the objectives were set, as follows:

1. Develop a reliable and practical method for determination of the chemical energy
content of wastewater, and identify a reliable surrogate of the chemical energy in
wastewater and in sludge
2. Construct chemical energy and mass balances for 4 actual wastewater treatment plants
(WWTPs) to determine the fate of chemical energy, organic carbon (expressed as COD),
nitrogen and phosphorus in the wastewater and sludge treatment (via modelling)
3. Identify potential waste recovery opportunities and explore the impact of their
implementation on the existing treatment process (via modelling)
8.1.1 Chemical energy of wastewater and sludge
The first objective of this study is to develop a reliable and practical method for determination
of the chemical energy content of wastewater and identify a reliable surrogate of the chemical
energy in wastewater and in sludge.
This research successfully developed a new wastewater drying method that is at least as good
as the proven method of freeze-drying for retaining the energy-containing material but is able
to shorten the drying time for 1 L of wastewater sample from 2-4 weeks to 3 days. Having
developed the drying method, 46 spot samples and 61 composite samples of raw wastewater,
20 composite samples of primary effluent, and 17 composite samples of secondary effluent
collected from four WWTPs were measured for chemical energy content (kJ/L) in this study.
Subsequently, this research conducted statistical analysis (correlation and Best Subset of
202
regression) to investigate the relationship between chemical energy (Eww, kJ/L) and commonly
analyzed wastewater variables. COD was found to be the best surrogate of the energy content
of wastewater. It has a ubiquitously (across the four WWTPs) strong linear relationship (r2=
77 -93%, p<0.05, excluding the spot raw) with Eww. An empirical mathematical relationship
between Eww and COD is found 15.8 kJ/g COD.
A similar statistical study was also conducted for understanding the relationship between
sludge chemical energy (Es, kJ/L) and the other commonly analyzed sludge variables, Strong
linear relationships (r2>85%, p<0.05) were found between the Es and sludge total solid (TSs),
sludge volatile solid (VSs), based on 87 sludge samples collected from various treatment
processes, such as primary settlement, activated sludge, anaerobic digestion, of four WWTPs .
The empirical mathematical relationships between Es and, TSs and VSs are found 17.1 kJ/g TSs
and 21.9 kJ/g VSs. Via assuming the wastewater and sludge share the same relationship
between COD and VS, the sludge COD was estimated. A similar relationship between chemical
energy and COD is found in both sludge and wastewater.
The objectives of developing a quick but reliable method of estimating wastewater energy
content using routinely analyzed variables has been fulfilled. Since the chemical energy of
wastewater and sludge can be estimated much more quickly via COD, TSs and/or VSs, an
energy audit of a treatment process or a treatment works can now be easily conducted. The
energy recovery performance can be better monitored. This also helps discover potential
opportunities for improvement in energy recovery.
balance
The second objective of this study is to construct chemical energy and mass balances for four
actual WWTPs to determine the fate of chemical energy, organic carbon (expressed as COD),
nitrogen and phosphorus in the wastewater and sludge treatment via modelling.
This study constructed, calibrated and validated models for the wastewater and sludge
treatment processes of four WWTPs with a range of population equivalents, wastewater
treatment processes, and sludge treatment processes. In all cases, the difference between the
203
predicted results from the observed results was no more than ± 15%. Hence, these predicted
results were used to build the chemical energy balance and COD, nitrogen and nitrogen mass
balance for the four WWTPs.
8.1.2.1 COD mass balance and chemical energy balance

The COD mass and chemical energy balance of the four WWTPs shows that:
• 69.7 ± 12.8% of wastewater COD and chemical energy flowing into the works can be
captured in the sludge and sent for sludge treatment via digestion,
• 24.8 ± 5.5% of wastewater COD and chemical energy flowing into the works is lost to
the atmosphere, potentially as CO2
• 7.5 ± 4.4 % of wastewater COD and chemical energy flowing into the works remains
in the final effluent and is discharged to the water bodies.
The COD mass balance and chemical energy balance of the four WWTPs suggest that majority
of the COD and chemical energy loading were captured during the wastewater treatment. Only
a small proportion of COD and chemical energy loading remain in the final effluent.
The COD mass balance and chemical energy balance of the sludge treatment of WWTP C
shows that,
• 21.3 % of the sludge COD and chemical energy ends in the dewatered digested sludge
• 31.1% of the sludge COD and chemical energy is retained in the filtrates and centrates
• 43.7% of the sludge COD and 38.4% of the sludge chemical energy is recovered as
CH4.
On this basis, approximately 20-30% of the chemical energy in raw wastewater has the
potential to be recovered as CH4. However, only 6-9% of the chemical energy in raw
wastewater can be potentially recovered as electricity. In this study, even though the chemical
energy within the raw wastewater of the investigated WWTPs is 5 - 14 times higher than its
electricity consumption, the WWTPs are not always electricity self-sustaining.
8.1.2.2 Nitrogen and Phosphorus mass balance

The phosphorus mass balance of the four WWTPs shows that:
• 68.2 ± 14.3% of wastewater phosphorus flowing into the works can be captured in the
204
sludge and sent for sludge treatment via digestion,
• 31.8 ± 14.3 % of wastewater phosphorus energy flowing into the works remains in the
final effluent and is discharged to the water bodies,
• No wastewater phosphorus flowing into the works is lost to the atmosphere.
The nitrogen mass balance of the four WWTPs shows that:

• For WWTPs A and D has nitrification in the activated sludge process,
o 35.9 ± 10.4% of wastewater nitrogen flowing into the works can be captured in
the sludge and sent for sludge treatment via digestion,
o 25.9 ± 11.3 % of wastewater nitrogen flowing into the works is lost to the
atmosphere, potentially as N2 or N2O,
o 38.3 ± 0.9 % of wastewater phosphorus energy flowing into the works remains
in the final effluent and is discharged to the water bodies,
• For WWTPs B and C has nitrification in the activated sludge process,
o 33.0 ± 10.4% of wastewater nitrogen flowing into the works can be captured in
the sludge and sent for sludge treatment via digestion,
o No wastewater nitrogen flowing into the works is lost to the atmosphere,
potentially as N2 or N2O,
o 67.0 ± 15.3 % of wastewater phosphorus energy flowing into the works remains
in the final effluent and is discharged to the water bodies,
The nitrogen and phosphorus mass balance of the four WWTPs suggests that the approximately
two-thirds of the wastewater phosphorus and one-third of wastewater nitrogen are captured in
the sludge during the wastewater treatment. Unlike the COD and chemical energy, the final
effluent could contain a substantial proportion of 30-70% of the nutrient mass flowing into the
work. Moreover, this study finds nitrification may cause nitrogen lost to the atmosphere due to
the unexpected denitrification in the activated sludge process and the subsequent final
clarification.
The nitrogen and phosphorus mass balance of the sludge treatment study of the sludge
treatment of WWTP C shows that,
• 68.5% of the sludge nitrogen and 64.5% of the sludge nitrogen is distributed in the
filtrates and centrates of the sludge thickening and dewatering, and
205
• 31.5% of the sludge nitrogen and 35.5 % of the sludge nitrogen remain in the dewatered
digested sludge.
On this basis, approximately 20-30% of the nitrogen and 35-55% of the nitrogen flowing into
the works are estimated to end up in the filtrates and centrates which will be recirculated back
to the head of the work. Meanwhile, approximately 7-14% of the nitrogen and 19-30% of the
nitrogen flowing into the work are estimated to be kept in the digested sludge.
balance
The third objective of this study is to identify potential waste recovery opportunities and
explore the impact of their implementation on the existing treatment process (via modelling)
Based on the chemical energy and COD mass balance study, this study found that even though
the chemical energy within the raw wastewater of the investigated WWTPs is 5-14 times higher
than its electricity consumption, the WWTPs are not always electricity self-sustaining.
The energy recovery opportunities are 1) to increase the chemical energy feed to the anaerobic
digestion, 2) to recover the unrecovered chemical energy kept in the dewatered digested sludge
(biosolid), and 3) to recover energy from wastewater directly.
Therefore, according to the opportunities discovered, four measures are proposed

• improving the efficiency of solids removal in primary settlement
• reducing the sludge retention time prior to the digestion
• recovering energy from the dewatered digested sludge via pyrolysis, and
• reducing the electricity consumption by replacing the activated sludge processes with
energy-generating anaerobic membrane reactor (AnMBR) units.
The current nutrient recovery focused on the agricultural land application of the dewatered
digested sludge and fertilizer recovery of the phosphate and ammonium from the raw sludge
centrate and digested sludge centrate, based on the nitrogen and phosphorus balance studied,
this study considers the opportunity is to improve the capturing of phosphorus to the sludge in
order to increase the nutrient loading distrusted to sludge filtrates and centrates and digested
206
sludge. Thus, it proposed a measure of adding ferric dosing or enhanced biological phosphorus
removal (EBPR) to the current system.
The five proposed measures are built into case studies. The impact of their implementation on
the holistic energy balance and nutrient balance are explored mainly via modelling (exclude
the dewatered digested sludge pyrolysis) and are listed as followed:
• Improving the efficiency of solids removal in the primary settlement is able to increase the
feed chemical energy to the digester. Moreover, it also reduces the energy consumption on
the aeration process. Therefore, a better energy balance is achieved. However, the
uncertainty of this measure is that, if metal coagulant is used, it could bring adverse effect
to the digester. In nutrient wise, it the phosphate and ammonium loading distributed to the
raw sludge centrate and digested sludge centrate but increases total nitrogen and total
phosphorus loading in the digested sludge. Furthermore, it increases the nitrogen and
phosphorus loading distributed to the final effluent.
• Reducing the sludge retention time prior to the digestion reduces the chemical energy lost
to centrate and hence increaser the feed chemical energy to the digester. It delivers a better
energy balance. It also reduces the phosphate and ammonium loading distributed to the raw
sludge centrate and digested sludge centrate but increases total nitrogen and total
phosphorus loading in the digested sludge.
• Recovering energy from the dewatered digested sludge via pyrolysis leads to a better energy
balance since the chemical energy with dewatered digested sludge is unlocked. It makes no
impact on the phosphate and ammonium loading distributed to the raw sludge centrate and
digested sludge centrate.
• Replacing the activated sludge processes with energy-generating AnMBR units can lead to
a better energy balance if the AnMBR can be energy self-sufficient. That is because
substantial amount of the aeration energy is saved. However, this is highly uncertain
depending of the energy consumption (including both heating and pumping) on the process
itself (Martin et al., 2011). Meanwhile, this measure reduces both the phosphate and
ammonium loading distributed to the raw sludge centrate and digested sludge centrate and
207
the total nitrogen and total phosphorus loading in the digested sludge. More nitrogen and
phosphorus are therefore distributed to the final effluent.
• Implementation of both ferric dosing and EBPR to improve the phosphorus capturing barely
impacts the energy balance. The implementation of EBPR substantially improves the
phosphate loading distributed to the raw sludge centrate and digested sludge centrate but
barely impacts the phosphorus loading to the digested sludge. The implementation of ferric
dosing is in the opposite: it barely changes phosphate loading distributed to the raw sludge
centrate and digested sludge centrate but substantially increases the phosphorus loading to
the digested sludge. The nitrogen mass balance is only slightly impacted in this case.
The five case studies suggest that the measures aim to promote the energy balance is likely to
increase the nutrient loading distributed to the dewatered digested sludge or to the final effluent
but reduce the phosphate and ammonium loading distributed to the raw sludge centrate and
digested sludge centrate. This is adverse to the fertilizer recovery from the raw sludge and
digested sludge centrate but provides an opportunity of recovering nutrient directly from
wastewater. However, the measures aim to promote the nutrient recovery from sludge treatment
barely impacts the energy balance
Since the biosolid pyrolysis potentially has the least impact to the upstream wastewater and
sludge treatment process and could potentially reduce the landbank stress, the technical and
economic feasibility of its implementation on to the existing energy recovery system of WWTP
C is studied. The economic evaluation considering the payback period, NPV and IRR suggest
that the installation of biosolid pyrolysis is not currently economically feasible. From the
sensitivity study, the capital expenditure (CAPEX), pyrolysis efficiency and chemical energy
content are the most influential parameters to economic evaluation. This study considers
economic feasibility could be achieved in the future if CAPEX becomes lowered due to the
technology development, and/or incentive granted will be elevated as pyrolysis is encouraged
by the government.
208
8.2 Recommendation
The breakdown of energy material in wastewater and sludge

This study has demonstrated that the chemical energy content of wastewater can be estimated
by multiplying its COD concentration by a value of 15.8 kJ/g COD. This 15.8 kJ/g COD only
represents the relationship between COD concentration and the total chemical energy contained
in wastewater, but not the energy content of a gram of COD because part of the chemical energy
is from the TKN (Scherson and Criddle, 2014). Therefore, understanding the energy content
of a gram of COD is a potential research area. So is recovering energy from TKN.
Wastewater and sludge are both mixtures. Their chemical energy is from different groups of
chemicals. This research did not investigate the breakdown of the energy contribution from
each chemical as it was not a primary aim of the research. Nevertheless, that is an area
recommended for future research because it may help target energy recovery technologies to
specific energy containing materials that are not recovered using current methods.
The establishing model of the innovative value recovery technology

In the nutrient recovery, a substantial proportion of nutrient remains in the final effluent
discharged. Recovery opportunity is, therefore, to recover this nutrient directly from
wastewater. However, the modelling software does not contain such a unit of ion exchange or
algal reactor. And hence, the modelling of the nutrient recovery directly from wastewater is
therefore hampered. Moreover, the microbial fuel/electrolysis cell is another promising energy
generating biological treatment (Stoll et al., 2018, Cotterill et al., 2018). Since it produces
different products of H2 or electricity, it would be useful to model its implementation and then
compare to the CH4 generating anaerobic treatment. Therefore, it would be useful to has such
innovative recovery technologies programmed into the modelling software.
Chemical energy recovery

Although improving the solid removal efficiency of primary settlement is found beneficial to
the energy balance, the uncertainty of this measure is how to improve the solid removal
efficiency. If metal coagulant is used and is in high dosage, harmful effect may be caused to
the digester (Diamantis et al., 2013). Therefore, it is worth to discover effective coagulant for
improving the solid settling without harming the downstream digestion.
209
As mentioned in Section 5.2.2, WWTP B has a high percentage of chemical energy captured
in the SAS. The reason of it is considered as that the high ferric content of the inflowing
wastewater causes bio-flocculation which reduce the chemical energy loss due to COD
oxidation. Therefore, coagulant dosing in the activated sludge process could be a potential
approach to improve the capturing of chemical energy in the activated sludge.
As stated in Section 8.1.2.1, approximately 30% of the sludge energy entered into the sludge
treatment ended in the filtrates and centrates. This study suggested reducing the sludge
retention time to reduce the chemical energy loading. This measure mainly focuses on reducing
the soluble energetic COD formed from the hydrolysis, fermentation and digestion of the
sludge. However, as shown in Table 5-4, majority of the COD in the primary sludge filtrate and
in the SAS filtrate are pCOD. Therefore, improving the solid retaining efficiency of the sludge
thickening and dewatering process could also leads to greater feed chemical energy to the
digestion.
Moreover, apart from reducing the energy loading in the filtrates and centrates for promoting
energy production, effort could be made on recovering the chemical energy directly from the
filtrates and/or centrates. As shown in Table 5-4, the COD content of the filtrates varies from
1,000-10,000 mg/L. Such a high concentration is favored by the anaerobic treatment or
microbial fuel cells (Stoll et al., 2018). Therefore, it is worth to investigate the treatability of
the filtrates and centrates in the anaerobic treatments and energy balance.
The recovery of nutrient and energy directly from wastewater requires a concentrated feed
wastewater (Stoll et al., 2018, Münch and Barr, 2001). The measure for promoting value
recovery proposed in this study mainly focus on modifying the process inside the WWTPs and
is unable to change to concentration of the valuable COD, nitrogen and phosphorus in the raw
wastewater. Therefore, effort could potentially be made on recovering the valuable from its
source via decentralized treatment (Capodaglio, 2017).
210
Appendix A. TN Adjustment for Sludge Sample
In this study, Total Nitrogen content (TN%) was measured by the scientific service of Scottish
Water. Due to the logistical constraints the sludge was dried prior to the test. However,
Ammonia can be lost during the sludge drying (Maurer and Müller, 2012, Huett, 1997), and
hence leads to the lower TN% of the sludge. Therefore, this study firstly adjusts the TN% of
the sludge.
TN% of the feed and thickened or dewatered sludge of WWTP C’s thickening and dewatering
facilities was adjusted via Equation 3-8 (shown in Section 3.2.2.4). The adjusted TN% (as
TN%Adjusted) and the experimentally measured TN% (as TN%Measured) were compared. Results
are shown in Table A-1.
Table A-1 Adjusted TN% of different type of sludge of Howdon WWTP
No. of
Type of Sludge TN%Measured, % TN%Adjusted, % TN%Measured:TN%Adjusted
Sampling
Primary Sludge 3 2.66 ±0.46 3.18 ±0.68 1.19 ±0.08
Thickened Primary Sludge 3 2.58 ±0.25 2.90 ±0.67 1.12 ±0.06
SAS 4 5.27 ±0.60 5.88 ±0.64 1.12 ±0.01
Thickened SAS 4 6.67 ±0.78 6.75 ±0.79 1.01 ±0.00
Raw Sludge 5 3.60 ±1.24 4.58 ±1.27 1.28 ±0.12
Raw Centrifuged Cake 5 3.67 ±0.56 3.82 ±0.55 1.05 ±0.01
Digested Sludge 5 4.63 ±1.55 8.62 ±1.87 1.89 ±0.37
Digested Cake 5 4.65 ±0.20 5.21 ±0.30 1.12 ±0.02
In Chapter 4 to 6, the TN% used is adjusted according to the TN%Measured: TN%Adjusted of the
corresponding type of sludge. For WWTP A and D, even though their primary sludge was
mixed with SAS and humus, since the TN%Measured: TN%Adjusted of the primary sludge and
SAS is close the TN% of their primary sludge is adjusted by just multiplying by 1.19, which
is the TN%Measured: TN%Adjusted of primary sludge.
211
Appendix B. The value range of the kinetic and stoichiometry of the
biological models.
Table B-1 The value range of the kinetic and stoichiometry of the biological models
Parameter Defaul Lower Upper
Value boundary Boundary
Model Stoichiometry
Heterotrophic yield on soluble substrate
Aerobic heterotrophic yield on soluble substrate 0.6250 0.38001 0.75001
Anoxic heterotrophic yield on soluble substrate 0.5330 0.4264 0.62502
Methylotrophic Biomass
Aerobic methylotroph yield on methanol 0.4500 0.3600 0.5400
Anoxic methylotroph yield on methanol 0.3600 0.2880 0.4320
Fermentative Biomass
Yield of fermentative biomass 0.1800 0.1440 0.2160
Ammonia-Oxidizing Biomass
Ammonia-oxidizer yield 0.1800 0.05003 0.30003
Nitrite-Oxidizing Biomass
Nitrite-oxidizer yield 0.0600 0.0480 0.0720
Anammox Biomass
Biomass yield on NH4-N 0.1675 0.1340 0.2010
Poly-Phosphate-Accumulating Biomass (PAOs)

Aerobic yield on PAO growth 0.6250 0.5000 0.7500
Anoxic yield on PAO growth 0.5110 0.4088 0.62502
PHA storage yield 0.4000 0.10004 0.60004
Xpp storage yield 0.2000 0.10004 0.30004
Acetogenic Biomass
Acetogenic yield on propionate 0.0400 0.0320 0.0480
Hydrogenotrophic Methanogenic Biomass

Methanogenic yield on H2 0.0600 0.0480 0.0720
212
Acetoclastic Methanogenic Biomass
Methanogenic yield on acetate 0.0500 0.0400 0.0600
Unbiodegradable fraction from biomass decay

Unbiodegradable fraction from cell decay 0.1000 0.0800 0.25005
Soluble inert COD fraction

Fraction of inert COD during slowly biodegradable organic 0.0000 0.0000 0.0000
hydrolysis
Fraction of inert COD during inert residue hydrolysis 0.0000 0.0000 0.0000
Fraction of inert COD during inert organic hydrolysis 0.0000 0.0000 0.0000
Kinetics
Absorption of colloidal COD
Specific adsoprtion rate 0.1000 0.0800 0.1200
Saturation/inhibition coefficient for Xs/Xbh 0.0500 0.0400 0.0600
Heterotrophic biomass
Maximum specific growth rate on substrate 6.0000 0.60001 13.20001
Saturation/inhibition coefficient for SS 4.0000 3.20005 15.0000
Saturation coefficient for oxygen 0.2000 0.0100 0.50001
Saturation coefficient for nitrogen as nutrient 0.0500 0.04001 0.2000
Saturation coefficient for Nox-N as nutrient 0.5000 0.10003 0.55003
Saturation coefficient for phosphorus (nutrient) 0.0100 0.00504 0.01504
Saturation/inhibition coefficient for Sac 4.0000 3.50004 4.50004
Saturation/inhibition coefficient for Spro 4.0000 3.2000 4.8000
Reduction factor for denitrification on nitrate-N 0.3200 0.2560 0.90004
reduction factor for denitrification on nitrite-N 0.4800 0.3840 0.5760
Saturation coefficient for nitrite 0.1000 0.0800 0.1200
Saturation coefficient for nitrate 0.5000 0.4000 0.6000
Oxygen inhibition coefficient for denitrification 0.2000 0.1600 0.2400
Aerobic heterotrophic decay rate 0.4000 0.0200 1.60005
Anoxic reduction factor fro decay rate 0.9000 0.7200 1.0800
Anaerobic reduction factor for decay rate 0.6000 0.4800 0.7200
Methylotrophic Biomass
Maximum growth rate fro methylotrophs 1.3000 1.0400 1.5600
Methanol saturation coefficient for methyltrophs 0.5000 0.4000 0.6000
Saturation coefficient of nitrite for methyltrophs 0.1000 0.0800 0.1200
213
Saturation coefficient of nitrate fro methyltrophs 0.1000 0.0800 0.1200
Oxygen saturation for methyltrophs 0.2000 0.1600 0.2400
Reduction factor for denitrification on nitrite-N 0.6000 0.4800 0.7200
Oxygen inhibition ocefficient for denitrification 0.2000 0.1600 0.2400
Aerobic methylotrophic decay rate 0.2000 0.1600 0.2400
Anoxic recution factor fro decay rate 0.9000 0.7200 1.0800
Anearobic methylotrophic decay rate 0.6000 0.4800 0.7200
Ammonia-Oxidizing Biomass
Maximum growth rate for ammonia oxidizer 1.0000 0.20007 1.2000
Ammonia saturation coefficient for ammonia oxidizer 1.0000 0.20001 1.50001
Oxygen saturation for ammonia oxidizer 0.5000 0.20002 3.00001
Inhibition coefficient of FA for ammonia oxidizer 50.0000 40.0000 60.0000
Inhibition coefficient of FNA for ammonia oxidizer 0.2000 0.1600 0.2400
1
ammonia oxidizer aerobic decay rate 0.1500 0.0200 0.23006
anoxic reduction factor for decay rate 0.5000 0.4000 1.50001
anaerobic recution factor for decay rate 0.3000 0.2400 0.3600
Nitrite-Oxidizing Biomass
Maximum growth rate for nitrite oxidizer 1.0000 0.8000 1.2000
2
NItrite saturation coefficient for nitrite oxidizer 0.5000 0.0400 0.6000
2
Oxygen saturation for nitrite oxidizer 0.6800 0.1000 0.8160
Inhibition coefficient of FA for nitrite oxidizer 1.0000 0.8000 1.2000
Inhibition coefficient of FNA for nitrite oxidizer 0.0900 0.0720 0.1080
nitrite oxidizer aerobic decay rate 0.1700 0.1360 0.2040
Anammox Biomass
Maximum growth rate for anammox bacteria 0.0186 0.0148 0.0223
Ammonia saturation coefficient for anammox bacteria 0.7300 0.5840 0.8760
Nitrite saturation for anammox bacteria 0.5000 0.4000 0.6000
Oxygen saturation/inhibition for anammox bacteria 0.1000 0.0800 0.1200
aerobic decay rate of anammox bacteria 0.0058 0.0046 0.0070
Poly-Phosphate-Accumulating Biomass (PAOs)
214
Rate constant for storage of PHA 3.0000 2.00004 6.00004
Saturation coefficient of PAO for Sac 4.0000 2.00002 4.8000
Saturation coefficient for Xpp/Xbp 0.0100 0.0080 0.02004
Saturation coefficient of PAO for Spro 4.0000 3.2000 4.8000
4
Maximum growth rate of PAO 1.0000 0.6700 1.2000
Saturation coefficient for PHA 0.0100 0.00704 0.01504
Saturation coefficient for oxygen 0.2000 0.1600 0.2400
4
Rate constant for storage of poly-phosphate 1.5000 1.0000 2.50002
Maximum ratio of Xpp/Xpao 0.3400 0.20008 0.51008
Inhibition coefficient for Xpp/Xbp 0.0200 0.01004 0.03004
P saturation for uptake 0.2000 0.10004 0.30004
Reduction factor for denitrification on nitrite-N 0.3600 0.2880 0.4320
Saturation coefficient of nitrite for PAO 0.5000 0.4000 0.6000
Saturation coefficient of nitrate for PAO 0.5000 0.4000 0.6000
Oxygen inhibition coefficient for denitrification 0.2000 0.1600 0.2400
4
aerobic decay rate for PAO 0.2000 0.1000 0.2400
2
anaerobic reduction factor for decay rate 0.6000 0.4800 0.7200
4
Poly-P lysis coefficient 0.2000 0.1000 0.30004
PHA lysis coefficient 0.2000 0.10004 0.2400
Fermentative Biomass
Maximum fermentation rate 3.0000 1.50004 3.6000
Oxygen saturation for obligate anaerobic biomass 0.1000 0.0800 0.1200
Nitrate saturation forobligate anaeobic biomass 0.1000 0.0800 0.1200
Substrate saturation for fermentative biomass 4.0000 3.2000 4.8000
Hydrogen saturation/inhibition for acidifier 10.0000 8.0000 12.0000
Aerobic decay rate for fermentative biomass 0.1333 0.1066 0.1600
Anoxic reduction factor for decay rate 0.5000 0.4000 0.6000
anaerobic reduction factor for decay rate 0.3000 0.2400 0.3600
Acetogenic Biomass
Maximum growth rate of propionate degrading bacteria 0.3500 0.2800 0.4200
Undissociated propionate saturation for propionate 10.0000 8.0000 12.0000
degrading bacteria
Hydrogen inhibition for propionate degrader 5.0000 4.0000 6.0000
Aerobic decay rate for acetogens 0.0670 0.0536 0.0804
215
Hydrogenotrophic Methanogenic Biomass
Maximum growth rate of H2-utilizing bacteria 0.3680 0.2944 0.4416
Hydrogen saturation for hydrogenotrophic methanogens 2.5000 2.0000 3.0000
Aerobic decay rate for hydrogenotrophic methanogens 0.0330 0.0264 0.0396
Acetoclastic Methanogenic Biomass

Maximum growth rate of acetate utilizing bacteria 0.1500 0.1200 0.1800
Acetate saturation for hydrogenotrophic methanogens 75.0000 60.0000 90.0000
Aerobic decay rate for acetoclastic methanogens 0.0670 0.0536 0.0804
Hydrolysis
Hydrolysis rate constant for xs 3.0000 0.96001 3.6000
7
Saturation coefficient for particulate COD 0.1000 0.0100 0.20002
Anoxic hydrolysis reduction factor 0.6000 0.40005 1.00005
Anaerobic hydrolysis reduction factor 0.4000 0.10002 0.45004
Saturation /ihibiton coefficient for NOx 0.5000 0.4000 0.6000
Hydrolysis rate constant for inert residue 0.0300 0.0240 0.0360
Saturation coefficient for inert residue 1.0000 0.8000 1.2000
Hydrolysis rate constant for inert organic 0.0300 0.0240 0.0360
Saturation coefficient for inert organic 1.0000 0.8000 1.2000
Ammonification
Ammonification rate 0.0800 0.0640 0.0960
1: The value is reported in Jeppsson (1996)
2: The value is reported in Drewnowski et al. (2018)
3: The value is reported in Weijers and Vanrolleghem (1997)
4: The value is reported in Henze et al. (2000)
5: The value is recommended by GPS-X 7.0
6: The value is reported in Liwarska-Bizukojc et al. (2011)
7: The value is reported in Petersen et al. (2002)
8: The value is reported in Rieger et al. (2001)
216
Appendix C. Sludge density calculation
Sludge consists water, fix solid (FS) and volatile solid (VS). The latter two compose the total
solid (TS). Thus, both mass or volume of sludge is the total mass or volume of the three. As
volume is calculated by diving the mass (m) with specific density, thus
Equation C-1
𝑚𝑆𝑙𝑢𝑑𝑔𝑒 𝑚𝑊𝑎𝑡𝑒𝑟 𝑚𝑉𝑆 𝑚𝐹𝑆
= +( + )
𝜌𝑆𝑙𝑢𝑑𝑔𝑒 𝜌𝑊𝑎𝑡𝑒𝑟 𝜌𝑉𝑆 𝜌𝐹𝑆
where mSludge, mWater, mVS and mFS are the mass of the sludge, water content, volatile solid and
fixed solid, respectively; ρSludge, ρWater, ρVS and ρFS are the density of the sludge, water content,
volatile solid and fixed solid, respectively
By diving the mSludge on both side of Equation B-1, gets

Equation C-2
1 𝑚𝑊𝑎𝑡𝑒𝑟 /𝑚𝑆𝑙𝑢𝑑𝑔𝑒 𝑚𝑉𝑆 /𝑚𝑆𝑙𝑢𝑑𝑔𝑒 𝑚𝐹𝑆 /𝑚𝑆𝑙𝑢𝑑𝑔𝑒
= +( + )
where 𝒎𝑾𝒂𝒕𝒆𝒓 /𝒎𝑺𝒍𝒖𝒅𝒈𝒆 is the percentage water content of the sludge which is calculated by
subtracting the TS% from 100%, 𝒎𝑽𝑺 /𝒎𝑺𝒍𝒖𝒅𝒈𝒆 is the percentage VS content of the sludge
which can be calculated by multiplying the TS% with the VS%, 𝒎𝑭𝑺 /𝒎𝑺𝒍𝒖𝒅𝒈𝒆 is the
percentage FS content of the sludge which can be calculated by multiplying the TS% with (1-
VS%). Thus,
Equation C-3
1 1 − 𝑇𝑆% 𝑉𝑆% × 𝑇𝑆% (1 − 𝑉𝑆%) × 𝑇𝑆%
= +( + )
The specific density of water is 1 g/cm3, and the specific density of VS and FS are commonly
taken as 1 g/cm3 and 2.5 g/cm3 (Andreoli, C.V., et al, 2007), respectively. Therefore, the specific
density of sludge is
Equation C-4
1
𝜌𝑆𝑙𝑢𝑑𝑔𝑒 =
1 − 𝑇𝑆% 𝑉𝑆% × 𝑇𝑆% (1 − 𝑉𝑆%) × 𝑇𝑆%
+ ( + )
1 1 2.5
where TS% is the total solid percentage, VS% is the volatile solid percentage (to total solid).
217
Sewage sludge collected in this study commonly has TS% varies from 0.5% to 24.0%, and VS%
varies from 60% to 80%. In this range of TS% and VS%, the specific density of sludge
increases with the increase of TS% and the decrease of VS% but is within range of 1.001 to
1.061 ton/m3. Thus, the specific density of all the sludge in this study were assumed as 1 ton/m3
(1,000 g/L).
218
Appendix D. Model calibration
The most important parameters of influent fraction, kinetic and stoichiometric that influence
the modelling results, and are therefore subject to adjustment in the model for each WWTP, are
listed in the Tables D-1 to D-4. The table also presents the δjmsqr (that indicates the influence of
the parameter to the entire model), the adjusted value, the order of adjustment, and the category
each adjusted parameter belongs to.
Table D-1 The calibrated parameters of the wastewater treatment model of WWTP A
Adjusted Tuning
Ranking Category* Parameters δjmsqr
Value Order
1 ASM Aerobic heterotrophic yield on soluble substrate 1.682 0.4277 Step 2
2 ASM Maximum growth rate for ammonia oxidizer 1.083 1.1306 Step 3
3 ASM Aerobic heterotrophic decay rate 0.741 0.0282 Step 2
4 Pri-ASM Anoxic heterotrophic yield on soluble substrate 0.388 0.4680 Step 3
5 Inf Soluble inert fraction of total COD 0.350 0.1768 Step 1
6 Final-ASM Aerobic heterotrophic decay rate 0.282 0.6955 Step 2
7 ASM Anoxic heterotrophic yield on soluble substrate 0.257 0.4320 Step 3
8 Pri-Set Flocculant zone settling parameter 0.213 0.0012 Step 2
9 Final-ASM Saturation coefficient for phosphorus (nutrient) 0.196 0.0085 Step 2
10 Final-ASM Maximum specific growth rate on substrate 0.192 3.9115 Step 2
11 Final-ASM Aerobic heterotrophic yield on soluble substrate 0.182 0.7491 Step 2
12 Inf Readily biodegradable fraction of total COD 0.182 0.0501 Step 1
13 Pri-Set Maximum Vesilind settling velocity 0.181 263.1532 Step 2
14 Final-Set Flocculant zone settling parameter 0.180 0.0020 Step 2
15 Final-Set Maximum Vesilind settling velocity 0.166 233.7769 Step 2
16 Pri-ASM Reduction factor for denitrification on nitrate-N 0.154 0.3001 Step 3
17 Pri-ASM Maximum specific growth rate on substrate 0.154 6.2283 Step 3
18 Pri-ASM Saturation coefficient for nitrate 0.150 0.5717 Step 3
19 Pri-Set Maximum non-settleable solids 0.144 16.4233 Step 2
20 ASM Saturation coefficient for nitrate 0.115 0.4301 Step 3
21 Inf P content of soluble inert material 0.115 0.0063 Step 1
22 Final-ASM Anoxic heterotrophic yield on soluble substrate 0.115 0.4507 Step 3
23 ASM Reduction factor for denitrification on nitrate-N 0.109 0.2889 Step 3
ASM Ammonia oxidizer aerobic decay rate** 0.0239 Step 3

Inf Colloidal fraction of slowly biodegradable COD 0.0494 Step 1
Chemical dosage for sCOD removal (kg/d) 0.7410 Step 1
*:“Inf” stands for influent fraction, “Pri-ASM” stands for the biological model of the primary settlement, “Pri-
Set” stands for the settling model of the primary settlement, “ASM” stands for the biological model of the activated
sludge process, “Final-ASM” stands for biological model of the final clarification, “Final-Set” stands the settling
model of the final clarification, “V-sCOD” stands for the virtual chemical dosing for sCOD removal, and “SYS-
BMS” stands for the biochemical model settings in the system setting of the model.
**: Based on the modeller’s knowledge, Ammonia oxidizer aerobic decay rate was added in the step 3 calibration
in order to adjust the predicted NH4+-N concentration close to the observed value.
219
Table D-2 The calibrated parameters of the wastewater treatment model of WWTP B
Adjusted Tuning
Ranking Category* Parameters δjmsqr
Value Order
1 Pri-Set Maximum Settleable TSS Fraction 1.006 0.6892 Step 2
Aerobic heterotrophic yield on soluble
2 ASM 0.572 0.7499 Step 2
substrate
5 ASM ammonia oxidizer aerobic decay rate 0.296 0.1714 Step 3
Ammonia saturation coefficient for ammonia
7 ASM 0.233 1.0501 Step 3
oxidizer
Anoxic heterotrophic yield on soluble
10 ASM 0.148 0.5485 Step 2
substrate
12 ASM Unbiodegradable fraction from cell decay 0.114 0.1537 Step 2
13 ASM Anoxic reduction factor for decay rate 0.111 0.5796 Step 3
Colloidal fraction of slowly biodegradable
14 Inf 0.106 0.1307 Step 1
COD
15 ASM Ammonia-oxidizer yield 0.097 0.1708 Step 3
16 Pri-ASM Rate constant for storage of poly-phosphate 0.095 1.6220 Step 3
17 Inf N content in soluble inert material 0.082 0.0516 Step 3
18 ASM Oxygen saturation for ammonia oxidizer 0.070 0.6622 Step 3
V-sCOD Chemical dosage for virtual sCOD removal 2.7956 Step 1

Chemical dosage for virtual phosphate
V-P 1250.6702 Step 1
removal
model of the final clarification, “V-sCOD” stands for the virtual chemical dosing for sCOD removal, V-P stands
for the virtual chemical dosing for phosphate removal, and “SYS-BMS” stands for the biochemical model settings
in the system setting of the model.
220
Table D-3 The calibrated kinetics and stoichiometry of the wastewater treatment model of WWTP C
Ranking Parameters Adjusted Tuning
Category* δjmsqr
Value Order
2 ASM 0.324 0.3830 Step 3
substrate
6 ASM Hydrolysis rate constant for xs 0.144 2.8962 Step 3

Inf 0.0453 Step 1
COD
V-sCOD Chemical dosage for sCOD removal (kg/d) 5.3398 Step 1
SYS-BMS N content of heterotrophic biomass** 0.0593 Step 3
SYS-BMS N content of unbiodegradable biomass** 0.0651 Step 3
SYS-BMS P content of heterotrophic biomass** 0.0411 Step 3
SYS-BMS P content of unbiodegradable biomass** 0.0323 Step 3
**: In order to adjust the predicted total nitrogen and total phosphorus concentration of the surplus activated
sludge close to the observed value. The range given to the N content of heterotrophic biomass, N content of
unbiodegradable biomass, P content of heterotrophic biomass, P content of unbiodegradable biomass were 0.05-
0.087 gN/g, 0.035-0.087 gN/gCOD, 0.015-0.056 gP/gCOD and 0.011-0.056 gP/COD, respectively. The range is
determined from the value reported in Brdjanovic et al. (2015) and Labelle et al. (2017)
The parameter of influent fraction (of digester feed sludge), kinetic and stoichiometric
parameters that influence the biological modelling results of the sludge storage, sludge
thickening and dewatering processes, and the anaerobic digestion, and are therefore subject to
adjustment in the model for each WWTP, are listed in the Tables D-5 to D-8. The table also
presents the adjusted value and the category those adjusted parameter belongs to.
221
Table D-4 The calibrated kinetics and stoichiometry of the wastewater treatment model of WWTP D
Ranking Parameters δjmsqr Adjusted Tuning
Category*
Value Order
1 ASM 0.486 0.7341 Step 2
substrate
Anoxic heterotrophic yield on soluble
3 ASM 0.310 0.5427 Step 2
substrate
6 ASM Ammonia oxidizer aerobic decay rate 0.198 0.2068 Step 3
Ammonia saturation coefficient for ammonia
7 ASM 0.184 1.0525 Step 3
oxidizer
Reduction factor for denitrification on nitrite-
9 ASM 0.168 0.5737 Step 3
N
Oxygen inhibition coefficient for
12 ASM 0.111 0.2400 Step 3
denitrification
13 Inf N content of soluble inert material 0.105 0.1500 Step 3
14 Inf 0.105 0.1508 Step 1
COD
16 ASM Oxygen saturation for ammonia oxidizer 0.097 0.2023 Step 3
19 ASM Unbiodegradable fraction from cell decay 0.080 0.1841 Step 2
20 ASM Maximum growth rate for nitrite oxidizer 0.075 0.8041 Step 3
V-sCOD Chemical dosage for sCOD removal (kg/d) 0.9992 Step 1

222
Table D-5 The calibrated kinetics and stoichiometry of the primary sludge storage, thickening and dewatering
Category* Parameters Adjusted Value
I-ASM Yield of fermentative biomass 0.1441
I-ASM Aerobic heterotrophic decay rate 0.8848
I-ASM Anaerobic reduction factor for decay rate (in Heterotrophic Biomass section) 0.6788
I-ASM Hydrolysis rate constant for xs 3.5989
I-ASM Anaerobic reduction factor for decay rate (in Hydrolysis section) 0.2680
PS-ASM Yield of fermentative biomass 0.1685

PS-ASM Aerobic heterotrophic decay rate 0.1498
PS-ASM Anaerobic reduction factor for decay rate (in Heterotrophic Biomass section) 0.5508
PS-ASM Hydrolysis rate constant for xs 3.3313
PS-ASM Saturation coefficient for particulate COD 0.1070
PS-ASM Anaerobic reduction factor for decay rate (in Hydrolysis section) 0.3913
PS-Drum Fraction of influent flow 0.5602

PS-Drum Removal efficiency of particulate inert material 0.9938
PS-Drum Removal efficiency of slowly biodegradable substrate 0.8141
PS-Drum Removal efficiency of nitrogen in slowly biodegradable substrate 0.7325
*: I-ASM and PS-ASM stands for the biological model of the virtual imported sludge tank and of the primary
sludge storage tank, respectively. Meanwhile, PS-Drum stands for the operational parameters of the drum
thickener of primary sludge.
Table D-6 The calibrated parameters of the SAS storage, thickening and dewatering
SAS-ASM Yield of fermentative biomass 0.1461
SAS-ASM Aerobic heterotrophic decay rate 0.4044
SAS-ASM Anaerobic reduction factor for decay rate (in Heterotrophic Biomass
0.5436
section)
SAS-ASM Hydrolysis rate constant for xs 0.9679
SAS-ASM Anaerobic reduction factor for decay rate (in Hydrolysis section) 0.1005
SAS-Drum Fraction of influent flow 0.1050

SAS-Drum Removal efficiency of particulate inert material 0.9980
SAS-Drum Removal efficiency of nitrogen in slowly biodegradable substrate 0.8143
SAS-Drum Removal efficiency of heterotrophic biomass 0.8198
*: SAS-ASM stands for the biological model of the virtual surplus activated sludge (SAS) tank. Meanwhile,
SAS-Drum stands for the operational parameters of the drum thickener of SAS.
223
Table D-7 The calibrated parameters of the raw sludge storage, thickening and dewatering
Raw-ASM Yield of fermentative biomass 0.1921
Raw-ASM Aerobic heterotrophic decay rate 0.4401
Raw-ASM Anaerobic reduction factor for decay rate (in Heterotrophic Biomass
0.5162
section)
Raw-ASM Hydrolysis rate constant for xs 2.2344
Raw-ASM Saturation coefficient for particulate COD 0.1179
Raw-ASM Anaerobic reduction factor for decay rate (in Hydrolysis section) 0.1778
Raw-Drum Fraction of influent flow 0.1643

Raw-Drum Removal efficiency of particulate inert material 0.9707
Raw-Drum Removal efficiency of slowly biodegradable substrate 0.9048
Raw-Drum Removal efficiency of nitrogen in slowly biodegradable substrate 0.9048
Raw-Drum Removal efficiency of phosphorus in slowly biodegradable substrate 0.9664
Raw-Drum Removal efficiency of heterotrophic biomass 0.9943
*: RAW-ASM stands for the biological model of the virtual raw sludge tank. Meanwhile, SAS-Drum stands for
the operational parameters of the centrifuge of raw sludge.
Table D-8 The calibrated parameters of the anaerobic digestion process and subsequent sludge storage,
thickening and dewatering

DF-Inf Particulate inert fraction of total COD 0.1994
DF-Inf N content of inert particulate material 0.0377
DF-Inf P content of inert particulate material 0.0263
AD-ASM Yield of fermentative biomass 0.2149

AD-ASM Fraction of inert COD during slowly biodegradable organic hydrolysis 0.0419
DS-Drum Fraction of influent flow 0.1296

DS-Drum Removal efficiency of particulate inert material 0.9983
DS-Drum Removal efficiency of unbiodegradable cell product 0.4952
DS-Drum Removal efficiency of fermentative biomass 0.9657
DS-Drum Removal efficiency of acetogenic biomass 0.6823
DS-Drum Removal efficiency of acetolactic methanogenic biomass 0.6823
DS-Drum Removal efficiency of hydrogenotrophic methanogenic biomass 0.7067
Sys-BMS C content of inert particulate material 0.3092

Sys-BMS C content of slowly biodegradable substrate 0.3167
*: DF-Inf stands for influent fraction of the digester feed, AD-ASM stands for the biological model of the
anaerobic digestion, DS-Drum stands for the operational parameters of the drum thickener of digested sludge,
and Sys-BMS the biochemical model settings in the system setting of the model.
224
Appendix E. The observed and predicted flow and concentration data
Table E-1 The observed and predicted flow and concentration of the modelling of wastewater treatment of four WWTPs
Stream Unit Observed Predicted Observed Predicted Observed Predicted Observed Predicted
Value Value Value Value Value Value Value Value
Primary Effluent sCOD, mg/L 193.4 200.1 193.2 187.4 180.8 180.7 86.5 87.1
Primary Effluent pCOD, mg/L 227.9 206.4 199.0 197.8 276.9 276.9 82.3 92.9
Primary Effluent NH4+-N, mg/L 27.6 29.4 34.6 37.7 43.1 41.8 20.8 19.5
Primary Effluent TKN, mg/L 34.7 38.9 42.0 46.6 58.6 51.3 27.9 26.7
Primary Effluent PO43--P, mg/L 1.8 2.3 4.9 5.3 7.1 6.7 1.6 1.5
Primary Sludge COD, mg/L 34844.1 34843.9 43951.2 43120.0 * * 24212.7 24210.8
Primary Sludge TN, mg/L 1125.7 945.8 1130.0 1336.0 * * 835.8 1013.8
Primary Sludge TP, mg/L 518.2 308.5 379.0 298.7 * * 243.2 264.4
Final Effluent sCOD, mg/L 21.3 19.6 29.0 24.6 47.0 46.5 23.5 21.9
Final Effluent pCOD, mg/L 19.9 19.8 16.8 13.9 46.8 46.2 5.0 5.8
Final Effluent NH4+-N, mg/L 0.20 0.88 30.1 31.9 37.9 38.0 2.3 3.0
Final Effluent TKN, mg/L 4.1 4.4 30.5 33.9 40.7 44.4 8.7 7.0
Final Effluent Nitrate-N, mg/L 13.4 13.7 0.0 0.0 0.0 0.0 2.9 4.7
Final Effluent TN, mg/L 17.5 20.6 30.5 33.9 40.7 44.4 11.8 12.2
Final Effluent Op, mg/L 0.9 1.0 3.0 2.1 3.8 3.8 1.6 1.4
Surplus Flow, m3/d 204.0 222.9 2712.4 2460.5 6041.0 6238.2 1032.0 888.7
Activated Sludge
Surplus COD, mg/L 6159.5 5425.4 5053.1 5195.8 6389.5 6949.5 2522.4 2809.3
Activated Sludge
Surplus TN, mg/L 377.9 348.1 284.0 287.3 358.0 351.6 163.6 188.8
Activated Sludge
Surplus TP, mg/L 99.8 96.1 121.0 89.9 183.0 179.1 39.2 43.4
Activated Sludge
*: No sample was collected, and hence no observed value was obtained.
230
Table E-2 The observed and predicted flow and concentration of the modelling of sludge thickening, dewatering and digestion of WWTP C
Primary sludge thickening SAS thickening Raw sludge dewatering Digested sludge dewatering
Observed Predicted Observed Predicted Observed Predicted Observed Predicted
Stream Unit
Value Value Value Value Value Value Value Value
Feed Sludge COD, mg/L 37917.7 37472.7 * * 41780.8 46870.5 49061.1 40834.3
Feed Sludge TN, mg/L 1051.0 1051.1 * * 1857.1 1705.0 4797.3 4315.0
Feed Sludge TP, mg/L 368.0 367.9 * * 629.2 691.4 1082.2 1099.1
Filtrate/Centrate COD, mg/L 9775.0 9780.6 1598.3 1609.3 10595.0 10582.5 6520.0 6471.3
Filtrate/Centrate sCOD,
2657.9 2648.4 34.9 47.7 6072.5 6062.9 4074.6 3993.1
mg/L
Filtrate/Centrate pCOD,
7117.1 7132.2 1563.4 1561.6 4522.5 4519.6 2445.4 2478.3
mg/L
Filtrate/Centrate NH4+-N,
164.5 164.6 32.9 40.6 428.8 491.3 2243.1 2238.7
mg/L
Filtrate/Centrate TKN, mg/L 273.5 273.9 88.5 88.5 530.3 597.8 3244.3 2608.7
Filtrate/Centrate PO43--P,
66.2 31.7 16.6 5.7 203.8 178.7 228.5 228.5
mg/L
Dewatered
COD, mg/L 59813.1 60021.3 47741.7 50480.8 244210.7 253699.5 276618.2 255423.9
Sludge
Dewatered
TN, mg/L 1646.4 1684.0 2972.1 2496.9 8127.4 8015.9 14991.3 14970.8
Sludge
Dewatered
TP, mg/L 547.0 504.4 1359.4 1359.6 2796.7 2796.6 5281.9 5879.1
Sludge
Biogas, m3/d m3/d * * * * * * 39633.1 39940.0
CO2, m3/d m3/d * * * * * * 15853.3 15500.0
CH4, m3/d m3/d * * * * * * 23779.9 24360.0
*: No sample was collected, and hence no observed value was obtained.
231
Appendix F. Conceptual modelling for investigating the impact of
population equivalent to energy and mass balance
Introduction
This case study is to investigate the impact of population equivalent on the COD, total nitrogen
(TN) and total phosphorus (TP) mass balance and the chemical energy balance.
Method
To a wastewater treatment plant (WWTP), the most noticeable impact of the population
equivalent is on the flow. Therefore, this case study investigates the chemical and energy
balance of four scenarios with different flow conditions of 10,000 m3/d, 50,000 m3/d, 150,000
m3/d, and 250,000 m3/d, respectively. The mass balance and chemical energy balance
construction is based on the predicted result from conceptual modelling.
A conceptual wastewater treatment work consists of primary settlement, activated sludge, and
final clarification are constructed. The primary sludge and surplus activated sludge (SAS) are
discharged separately. The conceptual model CM-F-1 is shown in Figure F-1.
Figure F-1 The conceptual model CM-F-1
The key features of the model are listed:

• For dimension wise, in order to provide the same extent of treatment, the hydraulic
retention time of the primary settlement, activated sludge, and final clarification are
232
assumed 3 hours, 2.4 hours, and 3 hours, respectively, according to the HRT of the
corresponding process of the four investigated WWTPs (calculated via tank size stated
in Table 3-2 and 3-3 and flow data stated in Table 5-1). Therefore, the tank sizes are
varied accordingly as shown in Table F-1
Table F-1 Dimensions of the primary settler, aeration tank and final clarifier
Primary Settler* Aeration Tank** Final Clarifier*
Flow, Side
Side Wall Centre Surface Depth Tank Centre Surface
m3/d Wall
Depth, m Depth, m Area, m2 ,m Volume, m3 Depth, m Area, m2
Depth, m
10000 2.0 3.6 500 4.0 2000 2 3.5 1000
50000 2.0 3.6 2500 4.0 10000 2 3.5 5000
150000 2.0 3.6 7500 4.0 30000 2 3.5 15000
250000 2.0 3.6 12500 4.0 50000 2 3.5 25000
*: The Sloping bottom circular primary clarifier is used in both the primary settlement and final clarification
**: The plug flow tank unit is used in the activated sludge.
• For control wise, in all four scenarios,

o The proportional–integral–derivative (PID) controller is applied to the
discharge of primary sludge. The total suspended solid (TSS) content of the
discharged primary sludge is the control variable and is set as 30,000 mg/L,
o The dissolved oxygen (DO) setpoint is 2.0 mg/L,
o The PID controller is applied to the discharge of SAS. The mixed liquor
suspended solid is the control variable and is set as 2,000 mg/L.
• For mechanical and biological reaction wise,

o The kinetics of the settling of the primary settlement and final clarification
remains the default value,
o Both primary settlement tank and final clarification tank are biological reactive
and are control by the same virtual “Plug Flow Tank” unit as shown in Figure
F-1,
o The stoichiometry and kinetics of the biological reactions are the same as the
value stated in the “Initial Value” column of Table B-1.
• The influent characterization of the raw wastewater remains the software default value,
expect the total phosphorus content and ortho-phosphate content are adjusted to 5.0
mg/L and 3 mg/L, respectively.
233
Results and discussions
The COD, TN and TP mass balance and the chemical energy balance of four scenarios four
scenarios with different flow conditions are constructed based on the predicted result and are
shown in Table F-2.
Table F-2 Chemical energy balance and COD, TN and TP mass balance of the conceptual treatment work in
different flow condition
Flow, Captured in Captured in Remained in
Subject Loss
m3/d Primary Sludge SAS Final Effluent
10,000 COD 30.3% 28.3% 8.3% 33.1%
50,000 COD 30.3% 29.9% 7.0% 32.8%
150,000 COD 30.3% 28.3% 8.3% 33.1%
250,000 COD 30.3% 28.4% 8.3% 33.1%
10,000 Chemical Energy 30.3% 28.3% 8.3% 33.1%

50,000 Chemical Energy 30.3% 29.9% 7.0% 32.8%
150,000 Chemical Energy 30.3% 28.3% 8.3% 33.1%
250,000 Chemical Energy 30.3% 28.4% 8.3% 33.1%
10,000 TN 14.3% 20.2% 46.3% 19.2%

50,000 TN 14.3% 21.5% 47.2% 17.1%
150,000 TN 14.3% 20.2% 46.3% 19.2%
250,000 TN 14.3% 20.3% 46.3% 19.2%
10,000 TP 16.8% 41.5% 41.7% 0.0%

50,000 TP 16.8% 43.7% 39.5% 0.0%
150,000 TP 16.8% 41.5% 41.8% 0.0%
250,000 TP 16.8% 41.6% 41.6% 0.0%
The data stated in Table F-2 shows that the mass balance and chemical energy balance are more
or less identical in different flow conditions. This suggest that, if the HRT, kinetic of the settling
process, stoichiometry and kinetic of biological reaction, and the quality of the influent
wastewater are the same, the mass balance and chemical energy balance will not be influenced
by the change of flow of the raw wastewater.
234
References
AGU, C. E., HJULSTAD, Å., ELSETH, G. & LIE, B. 2017. Algorithm with improved
accuracy for real-time measurement of flow rate in open channel systems. Flow
Measurement and Instrumentation, 57, 20-27.
AIC 2011. AIC Fertiliser Statistics Report 2011. Peterborough, UK: Agricultural Industries
Confederation.
AINGER, C., BUTLER, D., CAFFOR, I., CRAWFORD-BROWN, D., HELM, D. &
STEPHENSON, T. 2009. A Low Carbon Water Industry in 2050. Bristol, UK:
Environment Agency.
APHA 2012. Standard methods for the examination of water and wastewater, Washington, DC,
USA, Clearway Logistics Phase 1a.
BANKS, C. 2009. Optimising anaerobic digestion. Anaerobic digestion workshop. University
of Reading.
BASILE, A., CASSANO, A. & RASTOGI, N. K. 2015. Advances in membrane technologies
for water treatment: materials, processes and applications, Elsevier.
BATSTONE, D. J., KELLER, J., ANGELIDAKI, I., KALYUZHNYI, S., PAVLOSTATHIS,
S., ROZZI, A., SANDERS, W., SIEGRIST, H. & VAVILIN, V. 2002. The IWA
anaerobic digestion model no 1 (ADM1). Water Science and Technology, 45, 65-73.
BEIS 2017a. CHPQA case study - Coventry Energy from Waste. UK: Department for Business,
Energy & Industrial Strategy.
BEIS 2017b. CHPQA case study - Lister Hospital. UK: Department for Business, Energy &
Industrial Strategy.
BEIS 2017c. CHPQA case study - University of Liverpool. UK: Department for Business,
Energy & Industrial Strategy.
BEIS 2018. 2017 UK Provisional Greenhouse Gas Emissions. London, UK: Department for
Business, Energy & Industrial Strategy
BHOGAL, A., ROLLETT, A. & WILLIAMS, J. 2017. Use of food-based PAS110 digestate to
grow energy grasses on brownfield land as an AD feedstock. UK: WRAP.
BONMATı, A. & FLOTATS, X. 2003. Air stripping of ammonia from pig slurry:
characterisation and feasibility as a pre-or post-treatment to mesophilic anaerobic
digestion. Waste management, 23, 261-272.
BOUZAS, A., RIBES, J., FERRER, J. & SECO, A. 2007. Fermentation and elutriation of
primary sludge: Effect of SRT on process performance. Water Research, 41, 747-756.
BOWEN, A., EVANS, B., OLIVER, B., EVANS, R. & MERRY, J. Advanced Digestion at
Cardiff and Afan; Dwr Cymru Welsh Water drive for lowest sustainable cost of sludge
treatment and 15% reduction in carbon footprint. 15th Biosolids Conference, 2010.
BOYLES, W. 1991. The Science of CHEMICAL OXYGEN DEMAND. USA: Hach Company.
BRDJANOVIC, D., MEIJER, S. C., LOPEZ-VAZQUEZ, C. M., HOOIJMANS, C. M. & VAN
LOOSDRECHT, M. C. 2015. Applications of Activated Sludge Models, Iwa Publishing.
BRUN, R., KÜHNI, M., SIEGRIST, H., GUJER, W. & REICHERT, P. 2002. Practical
identifiability of ASM2d parameters—systematic selection and tuning of parameter
subsets. Water research, 36, 4113-4127.
235
BUNCE, J. T., NDAM, E., OFITERU, I. D., MOORE, A. & GRAHAM, D. W. 2018. A Review
of Phosphorus Removal Technologies and Their Applicability to Small-Scale Domestic
Wastewater Treatment Systems. Frontiers in Environmental Science, 6.
CAI, T., PARK, S. Y. & LI, Y. 2013. Nutrient recovery from wastewater streams by microalgae:
Status and prospects. Renewable and Sustainable Energy Reviews, 19, 360-369.
CAO, Y. & PAWŁOWSKI, A. 2012. Sewage sludge-to-energy approaches based on anaerobic
digestion and pyrolysis: Brief overview and energy efficiency assessment. Renewable
and Sustainable Energy Reviews, 16, 1657-1665.
CAPODAGLIO, A. 2017. Integrated, decentralized wastewater management for resource
recovery in rural and peri-urban areas. Resources, 6, 22.
CHERNICHARO, C. D. 2006. Post-treatment options for the anaerobic treatment of domestic
wastewater. Reviews in Environmental Science and Bio/Technology, 5, 73-92.
COLBERT, J., XIHENG, H. & KIRKLIN, D. 1981. Enthalpy of combustion of
microcrystalline cellulose. JOURNAL OF RESEARCH of the National Bureau of
Standards, 86.
COSENZA, A., MANNINA, G., VANROLLEGHEM, P. A. & NEUMANN, M. B. 2014.
Variance-based sensitivity analysis for wastewater treatment plant modelling. Sci Total
Environ, 470-471, 1068-77.
COTTERILL, S. E., DOLFING, J., CURTIS, T. P. & HEIDRICH, E. S. 2018. Community
Assembly in Wastewater-Fed Pilot-Scale Microbial Electrolysis Cells. Frontiers in
Energy Research, 6.
COTTERILL, S. E., DOLFING, J., JONES, C., CURTIS, T. P. & HEIDRICH, E. S. 2017. Low
Temperature Domestic Wastewater Treatment in a Microbial Electrolysis Cell with 1
m2
Anodes: Towards System Scale-Up. Fuel Cells.
DE VRIES, S. C., POSTMA, R., VAN SCHOLL, L., BLOM-ZANDSTRA, G., VERHAGEN,
J. & HARMS, I. 2017. Extractive Nutrient Recovery as a Green Option for Managing
Phosphorus in Sidestreams and Biosolids Netherland: Wageningen University &
Research
DECC 2008. CHP Finance. UK: Department of Energy & Climate Change.
DECC 2014. RHI Biomethane Injection to Grid Tariff Review. UK: Department of Energy &
Climate Change.
DEFRA 2012. Waste water treatment in the United Kingdom – 2012. London, UK: Department
for Environment, Food and Rural Affairs.
DEFRA 2018. Sewage sludge in agriculture: code of practice for England, Wales and Northern
Ireland. In: DEPARTMENT FOR ENVIRONMENT, F. R. A. A. E. A. (ed.). UK:
Department for Environment, Food & Rural Affairs and Environment Agency.
DEMIRBAS, A. 2016. Calculation of higher heating values of fatty acids. Energy Sources,
Part A: Recovery, Utilization, and Environmental Effects, 38, 2693-2697.
DENG, Q. 2014. Ammonia Removal and Recovery from Wastewater Using Natural Zeolite:
An Integrated System for Regeneration by Air Stripping Followed Ion Exchange.
Master, University of Waterloo.
DESLOOVER, J., DE CLIPPELEIR, H., BOECKX, P., DU LAING, G., COLSEN, J.,
VERSTRAETE, W. & VLAEMINCK, S. E. 2011. Floc-based sequential partial
nitritation and anammox at full scale with contrasting N2O emissions. Water Research,
45, 2811-2821.
236
DIAMANTIS, V., VERSTRAETE, W., EFTAXIAS, A., BUNDERVOET, B., SIEGFRIED,
V., MELIDIS, P. & AIVASIDIS, A. 2013. Sewage pre-concentration for maximum
recovery and reuse at decentralized level. Water Sci Technol, 67, 1188-93.
DONG, Y. 2015. ENZYME PRETREATMENT OF FATS, OIL AND GREASE FROM
RESTAURANT WASTE TO PROLONG DRAIN FIELD EFFECTIVENESS. Master,
Michigan State University.
DREWNOWSKI, J., ZABOROWSKA, E. & DE VEGA, C. H. Computer Simulation in
Predicting Biochemical Processes and Energy Balance at WWTPs. E3S Web of
Conferences, 2018. EDP Sciences, 03007.
DROSTE, R. L. 1997. Theory and practice of water and wastewater treatment, New York,
New York : J. Wiley.
EL-TEMTAMY, S. A. & GENDY, T. S. 2014. Economic evaluation and sensitivity analysis
of some fuel oil upgrading processes. Egyptian Journal of Petroleum, 23, 397-407.
EL SHEIKH, R., GOUDA, A., SALEM, A. & HENDY, I. 2016. Analysis and Characterization
of Wastewater Nitrogen Components for using in Wastewater Modeling and Simulation.
International Journal of Advanced Research in Chemical Science, 3.
EUROPEAN COMMISSION. 2019. Implementation of the Circular Economy Action Plan
[Online]. European Commission. [Accessed 06-02-2019].
EVANS, F. W. & SKINNER, H. A. 1959. The heat of combustion of acetic acid. Transactions
of the Faraday Society, 55, 260-261.
EVANS, T. D. Recovering ammonium and struvite fertilisers from digested sludge dewatering
liquors. Proc. IWA Specialist Conference: Moving Forward–Wastewater biosolids
sustainability, 2007.
FABIYI, M., MENNITI, A., SCHAUER, P., BRITTON, A. & GOEL, R. 2016. Modeling &
Operational Case Study of a Full Scale Phosphorus Recovery System Coupled with
WASSTRIP®: Factors to Consider in Model Development & Insights for Optimal P
Recovery. Proceedings of the Water Environment Federation, 2016, 1506-1521.
FALL, C., ESPINOSA-RODRIGUEZ, M., FLORES-ALAMO, N., VAN LOOSDRECHT, M.
& HOOIJMANS, C. 2011. Stepwise calibration of the activated sludge model no. 1 at
a partially denitrifying large wastewater treatment plant. Water Environment Research,
83, 2036-2048.
FATTAH, K. P. 2012. Assessing struvite formation potential at wastewater treatment plants.
International Journal of Environmental Science and Development, 3, 548.
FLORES-ALSINA, X., COMAS, J. & RODRÍGUEZ-RODA, I. 2010. Analysis of rising
sludge risk in Activated Sludge Systems: from operational strategies to clarifier design.
FOLADORI, P., ANDREOTTOLA, G. & ZIGLIO, G. 2010. Sludge reduction technologies in
wastewater treatment plants, IWA publishing.
FOLEY, J., DE HAAS, D., YUAN, Z. & LANT, P. 2010. Nitrous oxide generation in full-
scale biological nutrient removal wastewater treatment plants. Water Res, 44, 831-44.
GALLO, A. 2019. A Refresher on Internal Rate of Return [Online]. Available:
https://hbr.org/2016/03/a-refresher-on-internal-rate-of-return [Accessed 20 Dec 2018].
GANS, N., MOBINI, S. & ZHANG, X. 2007. Mass and energy balances at the Gaobeidian
wastewater treatment plant in Beijing, China. MSc
GASSO, K., MOUROT, G. & RAGOT, J. 2002. Environmental systems modelling and
237
diagnosis using a multiple model approach. Computational Intelligent Systems For
Applied Research. World Scientific.
GERNAEY, K., VANROLLEGHEM, P. & LESSARD, P. 2001. Modeling of a reactive
primary clarifier. Water Science and Technology, 43, 73-81.
GERNAEY, K. V., VAN LOOSDRECHT, M. C., HENZE, M., LIND, M. & JØRGENSEN,
S. B. 2004. Activated sludge wastewater treatment plant modelling and simulation:
state of the art. Environmental Modelling & Software, 19, 763-783.
GIMENEZ, J. B., ROBLES, A., CARRETERO, L., DURAN, F., RUANO, M. V., GATTI, M.
N., RIBES, J., FERRER, J. & SECO, A. 2011. Experimental study of the anaerobic
urban wastewater treatment in a submerged hollow-fibre membrane bioreactor at pilot
scale. Bioresour Technol, 102, 8799-806.
GOUVEIA, J., PLAZA, F., GARRALON, G., FDZ-POLANCO, F. & PENA, M. 2015. Long-
term operation of a pilot scale anaerobic membrane bioreactor (AnMBR) for the
treatment of municipal wastewater under psychrophilic conditions. Bioresour Technol,
185, 225-33.
GRIZZETTI, B. & BOURAOUI, F. 2006. Assessment of nitrogen and phosphorus
environmental pressure at European scale. EUR Report, 22526, 66.
GU, Y., LI, Y., LI, X., LUO, P., WANG, H., WANG, X., WU, J. & LI, F. 2017. Energy Self-
sufficient Wastewater Treatment Plants: Feasibilities and Challenges. Energy Procedia,
105, 3741-3751.
HEIDRICH, E., CURTIS, T. & DOLFING, J. 2010. Determination of the internal chemical
energy of wastewater. Environmental Science & Technology, 45, 827-832.
HENZE, M. & COMEAU, Y. 2008. Wastewater characterization.
HENZE, M., GUJER, W., MINO, T., MATSUO, T., WENTZEL, M. C., MARAIS, G. V. R.
& VAN LOOSDRECHT, M. C. M. 1999. Activated Sludge Model No.2d, ASM2D.
Water Science and Technology, 39, 165-182.
HENZE, M., GUJER, W., MINO, T. & VAN LOOSDRECHT, M. C. 2000. Activated sludge
models ASM1, ASM2, ASM2d and ASM3, IWA publishing.
HIGGINS, M. J. & MURTHY, S. 2006. Examination of Reactivation and Regrowth of Fecal
Coliforms in Centrifuge Dewatered, Anaerobically Digested Sludges. Water
Environment Research Foundation Report.
HUANG, M.-H., LI, Y.-M. & GU, G.-W. 2010. Chemical composition of organic matters in
domestic wastewater. Desalination, 262, 36-42.
HUETT, D. O. 1997. Fertiliser use effciency by containerised nursery plants 2. Nutrient
leaching. Australian Journal of Agricultural Research, 48, 259-265.
HULSBEEK, J. J. W., KRUIT, J., ROELEVELD, P. J. & VAN LOOSDRECHT, M. C. M.
2002. A practical protocol for dynamic modelling of activated sludge systems. Water
Science and Technology, 45, 127-136.
HYDROMANTIS 2015. Webinar: Modelling Anaerobic Digestion Processes in GPS-X.
Youtube.
HYDROMANTIS 2017. GPS-X Technical Reference (GPS-X version 7). Canada.
IRENA 2016. The Power to Change: Solar and Wind Cost Reduction Potential to 2025.
International Renewable Energy Agency.
ISMAIL, I. M., FAWZY, A. S., ABDEL-MONEM, N. M., MAHMOUD, M. H. & EL-
HALWANY, M. A. 2012. Combined coagulation flocculation pre treatment unit for
238
municipal wastewater. Journal of Advanced Research, 3, 331-336.
IVANOV, V., KUANG, S., STABNIKOV, V. & GUO, C. 2009. The removal of phosphorus
from reject water in a municipal wastewater treatment plant using iron ore. Journal of
Chemical Technology & Biotechnology, 84, 78-82.
IWA 2016. Water Utility Pathways in a Circular Economy. London, UK: International Water
Association.
JAFFER, Y., CLARK, T. A., PEARCE, P. & PARSONS, S. A. 2002. Potential phosphorus
recovery by struvite formation. Water research, 36, 1834-1842.
JARDIN, N., THÖLE, D. & WETT, B. 2006. Treatment of sludge return liquors: experiences
from the operation of full-scale plants. Proceedings of the Water Environment
Federation, 2006, 5237-5255.
JEPPSSON, U. 1996. Modelling aspects of wastewater treatment processes.
JI, B., YANG, K., ZHU, L., JIANG, Y., WANG, H., ZHOU, J. & ZHANG, H. 2015. Aerobic
denitrification: A review of important advances of the last 30 years. Biotechnology and
Bioprocess Engineering, 20, 643-651.
JIMENEZ, J., MILLER, M., BOTT, C., MURTHY, S., DE CLIPPELEIR, H. & WETT, B.
2015. High-rate activated sludge system for carbon management – Evaluation of crucial
process mechanisms and design parameters. Water Research, 87, 476-482.
JONES, R., MACGREGOR, J. & MURPHY, K. 1989. State estimation in wastewater
engineering: application to an anaerobic process. Environmental monitoring and
assessment, 13, 271-282.
JONES, S. B., MEYER, P. A., SNOWDEN-SWAN, L. J., PADMAPERUMA, A. B., TAN, E.,
DUTTA, A., JACOBSON, J. & CAFFERTY, K. 2013. Process design and economics
for the conversion of lignocellulosic biomass to hydrocarbon fuels: fast pyrolysis and
hydrotreating bio-oil pathway. Pacific Northwest National Lab.(PNNL), Richland, WA
(United States).
JUHÁSZ, L. 2011. Net present value versus internal rate of return. Economics & Sociology, 4,
46-53.
KHUNJAR, W., LATIMER, R., BILYK, K., ROHRBACHER, J. & PITT, P. 2013. Extractive
Nutrient Recovery as a Green Option for Managing Phosphorus in Sidestreams and
Biosolids [Online]. Available:
https://static1.squarespace.com/static/54806478e4b0dc44e1698e88/t/548619fae4b0af
073c8d429c/1418074618937/Khunjar-NutrientRecovery-29Oct13.pdf [Accessed 10-
10-2017].
KIM, D.-J. 2013. Effect of Aeration on Nitrous Oxide (N2O) Emission from Nitrogen-
Removing Sequencing Batch Reactors. Journal of Microbiology and Biotechnology, 23,
99-105.
KIM, D., YOUNG PARK, K. & YOSHIKAWA, K. 2017. Conversion of Municipal Solid
Wastes into Biochar through Hydrothermal Carbonization. In: HUANG, W.-J. (ed.)
Engineering Applications of Biochar. London, UK: IntechOpen Limited.
KIM, J. & BENJAMIN, M. M. 2004. Modeling a novel ion exchange process for arsenic and
nitrate removal. Water research, 38, 2053-2062.
KLEEMANN, R., CHENOWETH, J., CLIFT, R., MORSE, S., PEARCE, P. & SAROJ, D.
2015. Evaluation of local and national effects of recovering phosphorus at wastewater
treatment plants: Lessons learned from the UK. Resources, Conservation and Recycling,
105, 347-359.
239
KORTH, B., MASKOW, T., GÜNTHER, S. & HARNISCH, F. 2017. Estimating the Energy
Content of Wastewater Using Combustion Calorimetry and Different Drying Processes.
Frontiers in Energy Research, 5, 23.
KORYTÁROVÁ, J. & POSPÍŠILOVÁ, B. 2015. Evaluation of investment risks in CBA with
Monte Carlo method. Acta Universitatis Agriculturae et Silviculturae Mendelianae
Brunensis, 63, 245-251.
KRISTENSEN, P., FRIBOURG-BLANC, B. & NIXON, S. 2004. Outlooks on Nutrient
Discharges in Europe from Urban Waste Water Treatment Plants.
LABELLE, M.-A., DOLD, P. L., GADBOIS, A., DÉLÉRIS, S. & COMEAU, Y. 2017.
Activated Sludge Production Parameters and Nutrient Content of Organic Sludge
Components. Water Environment Research, 89, 51-61.
LANGERGRABER, G., RIEGER, L., WINKLER, S., ALEX, J., WIESE, J., OWERDIECK,
C., AHNERT, M., SIMON, J. & MAURER, M. 2004. A guideline for simulation
studies of wastewater treatment plants. Water Science and Technology, 50, 131-138.
LI, L., ZHANG, S., LI, G. & ZHAO, H. 2012. Determination of chemical oxygen demand of
nitrogenous organic compounds in wastewater using synergetic photoelectrocatalytic
oxidation effect at TiO2 nanostructured electrode. Analytica chimica acta, 754, 47-53.
LIBERTI, L., BOARI, G., PETRUZZELLI, D. & PASSINO, R. 1981. Nutrient removal and
recovery from wastewater by ion exchange. Water research, 15, 337-342.
LIPU, M. S. H. & JAMAL, T. 2013. Techno-economic Analysis of Solar Concentrating Power
(CSP) in Bangladesh. Int. J. Adv. Renew. Energy Res, 2, 750-762.
LIWARSKA-BIZUKOJC, E., OLEJNIK, D., BIERNACKI, R. & LEDAKOWICZ, S. 2011.
Calibration of a complex activated sludge model for the full-scale wastewater treatment
plant. Bioprocess Biosyst Eng, 34, 659-70.
LU, J. 2006. Optimization of anaerobic digestion of sewage sludge using thermophilic
anaerobic pre-treatment.
MACK, J. 2012. Accounting for material-specific inflation rates in life-cycle cost analysis for
pavement type selection. Transportation Research Record: Journal of the
Transportation Research Board, 86-96.
MANNINA, G., COSENZA, A., VANROLLEGHEM, P. A. & VIVIANI, G. 2011. A practical
protocol for calibration of nutrient removal wastewater treatment models. Journal of
Hydroinformatics, 13, 575-595.
MARIA, A. Introduction to modeling and simulation. Proceedings of the 29th conference on
Winter simulation, 1997. IEEE Computer Society, 7-13.
MARTIN, I., PIDOU, M., SOARES, A., JUDD, S. & JEFFERSON, B. 2011. Modelling the
energy demands of aerobic and anaerobic membrane bioreactors for wastewater
treatment. Environmental technology, 32, 921-932.
MAURER, C. & MÜLLER, J. 2012. Ammonia (NH3) emissions during drying of untreated
and dewatered biogas digestate in a hybrid waste‐heat/solar dryer. Engineering in Life
Sciences, 12, 321-326.
MAYER, B. K., BAKER, L. A., BOYER, T. H., DRECHSEL, P., GIFFORD, M., HANJRA,
M. A., PARAMESWARAN, P., STOLTZFUS, J., WESTERHOFF, P. & RITTMANN,
B. E. 2016. Total Value of Phosphorus Recovery. Environ Sci Technol, 50, 6606-20.
MCCARTY, P. L., BAE, J. & KIM, J. 2011. Domestic wastewater treatment as a net energy
producer–can this be achieved? : ACS Publications.
240
MEIJER, S. C. F., VAN LOOSDRECHT, M. C. M. & HEIJNEN, J. J. 2001. Metabolic
modelling of full-scale biological nitrogen and phosphorus removing WWTP's. Water
research, 35, 2711-2723.
MILLS, N., PEARCE, P., FARROW, J., THORPE, R., KIRKBY, N. & WATER, T. The
influence of heat balance on the economics of advanced anaerobic digestion processes.
16th Biosolids Conference: Organic Resources. Exhibition. Aquaenviro, Leeds, 2011.
MILLS, N., PEARCE, P., FARROW, J., THORPE, R. B. & KIRKBY, N. F. 2014.
Environmental & economic life cycle assessment of current & future sewage sludge to
energy technologies. Waste Manag, 34, 185-95.
MINITAB BLOG EDITOR. 2013. How to Interpret Regression Analysis Results: P-values and
Coefficients [Online]. Minitab. Available: http://blog.minitab.com/blog/adventures-in-
statistics-2/how-to-interpret-regression-analysis-results-p-values-and-coefficients
[Accessed 06-06-2018].
MINITAB INC. 2017a. Create a fitted line plot and regression model that go through the
origin [Online]. Minitab Inc. . Available: https://support.minitab.com/en-
us/minitab/18/help-and-how-to/modeling-statistics/regression/supporting-
topics/regression-models/create-models-that-go-through-the-origin/ [Accessed 06-06-
2018 2018].
MINITAB INC. 2017b. Model summary table for Fit Regression Model [Online]. Available:
https://support.minitab.com/en-us/minitab/18/help-and-how-to/modeling-
statistics/regression/how-to/fit-regression-model/interpret-the-results/all-statistics-
and-graphs/model-summary-table/ [Accessed].
MINITAB INC. 2017c. Should I use Bonett's method or Levene's method for 2 Variances?
[Online]. Minitab Inc. Available: https://support.minitab.com/en-us/minitab/18/help-
and-how-to/statistics/basic-statistics/supporting-topics/tests-of-proportions-and-
variances/bonett-s-method-or-levene-s-method/ [Accessed].
MOLINOS-SENANTE, M., HERNÁNDEZ-SANCHO, F., SALA-GARRIDO, R. &
GARRIDO-BASERBA, M. 2010. Economic Feasibility Study for Phosphorus
Recovery Processes. Ambio, 40, 408-416.
MÜNCH, E. V. & BARR, K. 2001. Controlled struvite crystallisation for removing phosphorus
from anaerobic digester sidestreams. Water research, 35, 151-159.
MUSVOTO, E., WENTZEL, M. & EKAMA, G. 2000. Integrated chemical–physical processes
modelling—II. simulating aeration treatment of anaerobic digester supernatants. Water
Research, 34, 1868-1880.
NADEN, P., BELL, V., CARNELL, E., TOMLINSON, S., DRAGOSITS, U., CHAPLOW, J.,
MAY, L. & TIPPING, E. 2016. Nutrient fluxes from domestic wastewater: A national-
scale historical perspective for the UK 1800-2010. Sci Total Environ, 572, 1471-1484.
NELSON, N. O., MIKKELSEN, R. L. & HESTERBERG, D. L. 2003. Struvite precipitation
in anaerobic swine lagoon liquid: effect of pH and Mg: P ratio and determination of
rate constant. Bioresource technology, 89, 229-236.
NIH. 2018a. Compound Summary for CID 222 [Online]. National Institute of Health. Available:
https://pubchem.ncbi.nlm.nih.gov/compound/222#section=Heat-of-Combustion
[Accessed 01-07-2018 2018].
NIH. 2018b. Compound Summary for CID 1049 [Online]. National Institute of Health.
Available: https://pubchem.ncbi.nlm.nih.gov/compound/1049#section=Heat-of-
Combustion [Accessed 01-07-2018 2018].
241
NIST. 2018a. Benzene [Online]. National Institute of Science and Standard. Available:
https://webbook.nist.gov/cgi/cbook.cgi?ID=C71432&Mask=2 [Accessed 01-07-2018
2018].
NIST. 2018b. Toluene [Online]. National Institute of Standards and Technology. Available:
https://webbook.nist.gov/cgi/cbook.cgi?ID=C108883&Type=HCOMBL#HCOMBL
[Accessed 01-07-2018 2018].
NOWAK, O., SVARDAL, K., FRANZ, A. & KUHN, V. 1999. Degradation of particulate
organic matter—A comparison of different model concepts. Water science and
technology, 39, 119-127.
OECD. 2017. Average annual hours actually worked per worker [Online]. Organisation for
Economic Co-operation and Development. Available:
https://stats.oecd.org/Index.aspx?DataSetCode=ANHRS [Accessed 14-10-2017 2017].
OFGEM 2018. Non-domestic Renewable Heat Incentives, Guidance Volume 1: Eligibility and
how to apply. In: MARKETS, T. O. O. G. A. E. (ed.). UK.
OFGEM. 2019a. Feed-In Tariff (FIT) rates [Online]. Available:
https://www.ofgem.gov.uk/environmental-programmes/fit/fit-tariff-rates [Accessed 16
Dec 2018].
OFGEM. 2019b. Tariffs and payments: Non-Domestic RHI [Online]. Available:
https://www.ofgem.gov.uk/environmental-programmes/non-domestic-rhi/contacts-
guidance-and-resources/tariffs-and-payments-non-domestic-rhi [Accessed].
OFWAT 2016. Water 2020: our regulatory approach for water and wastewater services in
England and Wales Appendix 2 Moving beyond waste - further evidence and analysis.
UK: Office of Water Services.
OGILVIE, D. 1998. National Study of the Composition of Sewage Sludge. Auckland, New
Zealand: The New Zealand Water and Wastes Association.
OSTARA 2010. Case Study York Wastewater Treatment Plant, City of York Vancover,
Canada: Ostara.
OWEN, W. F. 1982. Energy in wastewater treatment.
PAZ-FERREIRO, J., NIETO, A., MÉNDEZ, A., ASKELAND, M. & GASCÓ, G. 2018.
Biochar from biosolids pyrolysis: A review. International journal of environmental
research and public health, 15, 956.
PEARCE, D. W. & TURNER, R. K. 1990. Economics of natural resources and the
environment, JHU Press.
PEARCE, P., MILLS, N. & WINTER, P. 2014. Novel sludge management options for a large
UK water company. Water Practice and Technology, 9, 179-185.
PETERSEN, B., VANROLLEGHEM, P. A., GERNAEY, K. & HENZE, M. 2002. Evaluation
of an ASM1 model calibration procedure on a municipal–industrial wastewater
treatment plant. Journal of Hydroinformatics, 4, 15-38.
POST 2007. ENERGY AND SEWAGE UK: Parliamentary Office of Science and Technology.
POWER, C., MCNABOLA, A. & COUGHLAN, P. 2014. Development of an evaluation
method for hydropower energy recovery in wastewater treatment plants: Case studies
in Ireland and the UK. Sustainable Energy Technologies and Assessments, 7, 166-177.
PUCHONGKAWARIN, C., GOMEZ-MONT, C., STUCKEY, D. C. & CHACHUAT, B. 2015.
Optimization-based methodology for the development of wastewater facilities for
energy and nutrient recovery. Chemosphere, 140, 150-8.
242
PUIG, S., VAN LOOSDRECHT, M., COLPRIM, J. & MEIJER, S. 2008. Data evaluation of
full-scale wastewater treatment plants by mass balance. Water research, 42, 4645-4655.
RABONI, M., TORRETTA, V. & URBINI, G. 2013. Influence of strong diurnal variations in
sewage quality on the performance of biological denitrification in small community
wastewater treatment plants (WWTPs). Sustainability, 5, 3679-3689.
RASI, S. 2009. Biogas composition and upgrading to biomethane, University of Jyväskylä.
RAWLINSON, D., COVERDALE, S., OLIVER, B. & ORD, J. 2012. Howdon & Bran Sands
STWs using advanced anaerobic digestion to improve operational efficiency and
environmental sustainability in Northumbrian Water. UK Water Projects Online. 2012
ed. UK.
REDDY, M., LEENHEER, J. & MALCOLM, R. 1989. Elemental analysis and heat of
combustion of fulvic acid from the Suwannee River. In: Humic Substances in the
Suwannee River, Georgia: Interactions, Properties, and Proposed Structures. Open-
File Report 87-557, 1989. p 147-161, 4 fig, 4 tab, 11 ref.
RIBEIRO DA SILVA, M. A. V., RIBEIRO DA SILVA, M. D. M. C., LOBO FERREIRA, A.
I. M. C., SHI, Q., WOODFIELD, B. F. & GOLDBERG, R. N. 2013. Thermochemistry
of α-D-xylose(cr). The Journal of Chemical Thermodynamics, 58, 20-28.
RIEGER, L., GILLOT, S., LANGERGRABER, G., OHTSUKI, T., SHAW, A., TAKACS, I.
& WINKLER, S. 2012. Guidelines for using activated sludge models, IWA publishing.
RIEGER, L., KOCH, G., KÜHNI, M., GUJER, W. & SIEGRIST, H. 2001. The EAWAG Bio-
P module for activated sludge model No. 3. Water Research, 35, 3887-3903.
ROELEVELD, P. J. & VAN LOOSDRECHT, M. C. M. 2002. Experience with guidelines for
wastewater characterisation in The Netherlands. Water Science and Technology, 45,
77-87.
RYSANEK, A. & CHOUDHARY, R. 2013. Optimum building energy retrofits under technical
and economic uncertainty. Energy and Buildings, 57, 324-337.
SAN ONG, T. & THUM, C. H. 2013. Net present value and payback period for building
integrated photovoltaic projects in Malaysia. International Journal of Academic
Research in Business and Social Sciences, 3, 153.
SARKAR, U., DASGUPTA, D., BHATTACHARYA, T., PAL, S. & CHAKROBORTY, T.
2010. Dynamic simulation of activated sludge based wastewater treatment processes:
Case studies with Titagarh Sewage Treatment Plant, India. Desalination, 252, 120-126.
SARPARASTZADEH, H., SAEEDI, M., NAEIMPOOR, F. & AMINZADEH, B. 2007.
Pretreatment of municipal wastewater by enhanced chemical coagulation. International
Journal of Environmental Research, 1, 104-113.
SATO, M. 1990. Thermochemistry of the formation of fossil fuels. In R.J. Spencer and I.M.
Chou (Eds) Fluid-Mineral Interactions: A Tribute to HP Eugster. Geol. Soc. Am. Spec.
Publ. , 2, 271-283.
SATOH, H., NAKAMURA, Y., ONO, H. & OKABE, S. 2003. Effect of oxygen concentration
on nitrification and denitrification in single activated sludge flocs. Biotechnol Bioeng,
83, 604-7.
SCHAUM, C., LENSCH, D. & CORNEL, P. 2016. Evaluation of the energetic potential of
sewage sludge by characterization of its organic composition. Water Sci Technol, 73,
3072-9.
SCHERSON, Y. D. & CRIDDLE, C. S. 2014. Recovery of freshwater from wastewater:
upgrading process configurations to maximize energy recovery and minimize residuals.
243
Environmental science & technology, 48, 8420-8432.
SEDLAK, R. I. 1991. Phosphorus and nitrogen removal from municipal wastewater:
principles and practice, CRC press.
SHAKYA, R., ADHIKARI, S., MAHADEVAN, R., SHANMUGAM, S. R., NAM, H.,
HASSAN, E. B. & DEMPSTER, T. A. 2017. Influence of biochemical composition
during hydrothermal liquefaction of algae on product yields and fuel properties.
Bioresource Technology, 243, 1112-1120.
SHI, C. Y. 2011. Mass flow and energy efficiency of municipal wastewater treatment plants,
IWA Publishing.
SHIZAS, I. & BAGLEY, D. M. 2004. Experimental determination of energy content of
unknown organics in municipal wastewater streams. Journal of Energy Engineering,
130, 45-53.
SHOENER, B. D., BRADLEY, I. M., CUSICK, R. D. & GUEST, J. S. 2014. Energy positive
domestic wastewater treatment: the roles of anaerobic and phototrophic technologies.
Environ Sci Process Impacts, 16, 1204-22.
SIEGRIST, H., KREBS, P., BÜHLER, R., PURTSCHERT, I., RÖCK, C. & RUFER, R. 1995.
Denitrification in secondary clarifiers. Water Science and Technology, 31, 205-214.
SIN, G., GERNAEY, K. V., NEUMANN, M. B., VAN LOOSDRECHT, M. C. & GUJER, W.
2011. Global sensitivity analysis in wastewater treatment plant model applications:
prioritizing sources of uncertainty. Water Res, 45, 639-51.
SINGH, H. 2007. Interactions of milk proteins during the manufacture of milk powders. Le
Lait, 87, 413-423.
SMYTH, M., MINALL, R., STEAD, T. & WALKER, J. 2016. Optimising the energy yield
from anaerobic digestion through calorific value analysis: case studies from davyhulme
and seafield. 21st Biosolids Conference. Edinburgh, UK: Aqua Enviro.
SNF FLOERGER 2003. Sludge Dewatering. In: FLOERGER, S. (ed.). France.
SÖTEMANN, S. W., WENTZEL, M. C. & EKAMA, G. A. 2006. Mass balance-based plant-
wide wastewater treatment plant models–Part 4: Aerobic digestion of primary and
waste activated sludges. Water SA, 32, 297-306.
STOICA, A., SANDBERG, M. & HOLBY, O. 2009. Energy use and recovery strategies within
wastewater treatment and sludge handling at pulp and paper mills. Bioresource
technology, 100, 3497-3505.
STOLL, Z., DOLFING, J. & XU, P. 2018. Minimum Performance Requirements for Microbial
Fuel Cells to Achieve Energy-Neutral Wastewater Treatment. Water, 10, 243.
STRATFUL, I., SCRIMSHAW, M. D. & LESTER, J. N. 2001. Conditions influencing the
precipitation of magnesium ammonium phosphate. Water research, 35, 4191-4199.
SUNNER, S. & MÅNSSON, M. 1979. Combustion Calorimetry: Experimental Chemical
Thermodynamics, Elsevier.
SUTTON, P. M., MELCER, H., SCHRAA, O. J. & TOGNA, A. P. 2011. Treating municipal
wastewater with the goal of resource recovery. Water Science and Technology, 63, 25-
31.
TAKÁCS, I., PATRY, G. G. & NOLASCO, D. 1991. A dynamic model of the clarification-
thickening process. Water research, 25, 1263-1271.
TAS, D. O., KARAHAN, Ö., ÖVEZ, S., ORHON, D. & SPANJERS, H. 2009.
Biodegradability and denitrification potential of settleable chemical oxygen demand in
244
domestic wastewater. Water Environment Research, 81, 715-727.
TCHOBANOGLOUS, G., BURTON, F. L. & STENSEL, H. D. 2003. Wastewater engineering
treatment and reuse, Boston, US: McGraw-Hill Higher Education.
THORNTON, I., BUTLER, D., DOCX, P., HESSION, M., MAKROPOULOS, C.,
MCMULLEN, M., NIEUWENHUIJSEN, M., PITMAN, A., RAUTIU, R. &
SAWYER, R. 2001. Pollutants in urban waste water and sewage sludge. Luxembourg:
European Commission.
TSUZUKI, T. & HUNT, H. 1957. Heats of Combustion. VI. The Heats of Combustion of Some
Amino Acids. The Journal of Physical Chemistry, 61, 1668-1668.
UBAY-COKGOR, E., OKTAY, S., ZENGIN, G., ARTAN, N. & ORHON, D. 2005. Effect of
primary sludge fermentation products on mass balance for biological treatment. Water
Science and Technology, 51, 105-114.
VAN DER HOEK, J., DUIJFF, R. & REINSTRA, O. 2018. Nitrogen Recovery from
Wastewater: Possibilities, Competition with Other Resources, and Adaptation
Pathways. Sustainability, 10.
VAN DER HOEK, J. P., DE FOOIJ, H. & STRUKER, A. 2016. Wastewater as a resource:
Strategies to recover resources from Amsterdam’s wastewater. Resources,
Conservation and Recycling, 113, 53-64.
VANHOOREN, H., MEIRLAEN, J., AMERLINCK, Y., CLAEYS, F., VANGHELUWE, H.
& VANROLLEGHEM, P. A. 2003. WEST: modelling biological wastewater treatment.
Journal of Hydroinformatics, 5, 27-50.
VANWORTSWINKEL, L. & NIJS, W. 2010. Industrial Combustion Boilers. International
Energy Agency.
WAN, J., GU, J., ZHAO, Q. & LIU, Y. 2016. COD capture: a feasible option towards energy
self-sufficient domestic wastewater treatment. Scientific reports, 6, 25054.
WANG, L., MIN, M., LI, Y., CHEN, P., CHEN, Y., LIU, Y., WANG, Y. & RUAN, R. 2010.
Cultivation of green algae Chlorella sp. in different wastewaters from municipal
wastewater treatment plant. Applied biochemistry and biotechnology, 162, 1174-1186.
WANG, T., CAMPS-ARBESTAIN, M., HEDLEY, M. & BISHOP, P. 2012. Predicting
phosphorus bioavailability from high-ash biochars. Plant and Soil, 357, 173-187.
WEIJERS, S. R. & VANROLLEGHEM, P. A. 1997. A procedure for selecting best identifiable
parameters in calibrating activated sludge model no. 1 to full-scale plant data. Water
science and technology, 36, 69-79.
WETT, B., BUCHAUER, K. & FIMML, C. Energy self-sufficiency as a feasible concept for
wastewater treatment systems. IWA Leading Edge Technology Conference, 2007.
Singa-pore: Asian Water, 21-24.
WILLIAMS, A. T., ZITOMER, D. H. & MAYER, B. K. 2015. Ion exchange-precipitation for
nutrient recovery from dilute wastewater. Environmental Science: Water Research &
Technology, 1, 832-838.
WILSON, L. & BOUWER, E. 1997. Biodegradation of aromatic compounds under mixed
oxygen/denitrifying conditions: a review. Journal of Industrial Microbiology and
Biotechnology, 18, 116-130.
WU YANG, X., RUI LIU, J., LI GAO, S., DONG HOU, Y. & ZHEN SHI, Q. 1999.
Determination of combustion energies of thirteen amino acids. Thermochimica Acta,
329, 109-115.
WWT ONLINE. 2015. Northumbrian Water creates gas from sewage for the national grid
245
[Online]. Available: http://wwtonline.co.uk/news/northumbrian-water-creates-gas-
from-sewage-for-the-national-grid#.WeJlhFXyvIU [Accessed 10-10-2017].
YESHI, C., LENG, L. C., LI, L., YINGJIE, L., SENG, L. K., GHANI, Y. A. & LONG, W. Y.
2013. Mass flow and energy efficiency in a large water reclamation plant in Singapore.
Journal of Water Reuse and Desalination, 3, 402-409.
ZALAT, O. & ELSAYED, M. 2013. A study on microwave removal of pyridine from
wastewater. Journal of Environmental Chemical Engineering, 1, 137-143.
ŽIŽLAVSKÝ, O. 2014. Net present value approach: method for economic assessment of
innovation projects. Procedia-Social and Behavioral Sciences, 156, 506-512.
246

Dai Z 2019

Uploaded by

Copyright:

Available Formats

Dai Z 2019

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dai Z 2019

Uploaded by

Copyright:

Available Formats

Developing Energy and Nutrient Mass Balances to

Inform Value Recovery Options in Municipal

Thesis submitted to Newcastle University for the degree of Doctor of

Author: Ziye Dai

Chapter 1 Introduction ............................................................................................................... 1

1.1 Background and drivers ................................................................................................... 1

1.1.1 Wastewater and wastewater treatment ...................................................................... 1

1.1.2 Drivers for value recovery from wastewater ............................................................ 2

1.1.3 Current status of value recovery from wastewater ................................................... 4

1.2 Aims and Objectives ........................................................................................................ 6

1.3 Thesis Structure ............................................................................................................... 8

Chapter 2 Literature Review ...................................................................................................... 9

2.1 The surrogate of chemical energy in wastewater and in sludge ...................................... 9

2.1.1 Surrogate of wastewater chemical energy .............................................................. 10

2.1.2 Surrogate of sludge chemical energy ...................................................................... 13

2.2 Fate of valuables in wastewater treatment work ............................................................ 13

2.2.1 Fate of COD, N and P in primary settlement .......................................................... 15

2.2.2 Fate of COD, N and P in secondary settlement ...................................................... 16

2.2.3 Fate of COD, N and P in sludge treatment ............................................................. 19

2.2.4 COD, N and P mass balance of wastewater treatment plants ................................. 20

2.3.1 Wastewater Mass balance Constraints .................................................................... 21

2.3.2 Applying modelling for mass balance construction ................................................ 22

2.3.3 Applying modelling for studying waste recovery ................................................... 23

2.4 Research Questions ........................................................................................................ 24

Chapter 3 Methodology ........................................................................................................... 26

3.1 Sample and data collection methods .............................................................................. 26

3.1.1 Wastewater and sludge sampling ............................................................................ 26

3.1.2 Sample storage ........................................................................................................ 27

3.2 Chemical analysis .......................................................................................................... 29

3.2.1 Chemical analysis of wastewater ............................................................................ 29

3.2.2 Chemical analysis of sludge.................................................................................... 33

3.3 Statistical analysis .......................................................................................................... 37

3.4 Wastewater and sludge treatment modelling ................................................................. 38

3.4.1 Model construction ................................................................................................. 38

3.4.2 Influent characterization ......................................................................................... 46

3.4.3 Model calibration and validation ............................................................................ 53

3.5 Construction of mass and chemical energy balances ..................................................... 60

3.5.1 Calculation of mass loading .................................................................................... 60

3.5.2 Calculation of chemical energy loading.................................................................. 61

3.5.3 Flow balance ........................................................................................................... 62

3.5.4 Mass balance ........................................................................................................... 63

3.5.5 Chemical energy balance ........................................................................................ 65

3.5.6 Calculations related to the primary treatment ......................................................... 67

4.1 Chemical energy of wastewater ..................................................................................... 70

4.1.1 Method Improvement.............................................................................................. 70

4.1.2 Wastewater characteristics ...................................................................................... 71

4.2 Chemical energy of sludge ............................................................................................. 83

4.2.1 Sludge Characteristic .............................................................................................. 83

4.3 Chapter Summary .......................................................................................................... 94

5.1 Model calibration and validation ................................................................................... 96

5.1.1 Wastewater and sludge treatment models ............................................................... 96

5.3.1 Energy product of wastewater treatment .............................................................. 116

5.3.2 Energy product of sludge treatment ...................................................................... 117

5.3.3 Energy recovery potential from wastewater ......................................................... 119

5.4 Case studies of energy recovery .................................................................................. 126

5.5 Chapter summary ......................................................................................................... 143

6.1.2 Nitrogen and phosphorus mass balance of secondary biological treatment as

6.2.1 Nutrient product of wastewater treatment............................................................. 152

6.2.2 Nutrient product of sludge treatment .................................................................... 155

6.2.3 Nutrient recovery potential from wastewater ....................................................... 157