Aspen Plus Model of An Alkaline Electrolysis System For Hydrogen Production

international journal of hydrogen energy xxx (xxxx) xxx
Available online at www.sciencedirect.com
ScienceDirect
journal homepage: www.elsevier.com/locate/he
Aspen Plus model of an alkaline electrolysis

system for hydrogen production
Monica Sa
nchez a,b,*, Ernesto Amores a, David Abad a, Lourdes Rodrı́guez c,
Carmen Clemente-Jul b
a
Centro Nacional del Hidrogeno (CNH2), Prolongacion Fernando El Santo s/n, 13500, Puertollano, Ciudad Real, Spain
b
Departamento de Energı́a y Combustibles, E.T.S.I de Minas y Energı́a, Universidad Politecnica de Madrid (UPM), C/
Rı́os Rosas 21, 28003, Madrid, Spain
c
Universidad Europea de Madrid (UEM), C/Tajo s/n, Urbanizacion El Bosque, 28670 Villaviciosa de Odon, Madrid,
Spain
highlights
A model of an alkaline electrolysis (AEL) plant has been developed using Aspen Plus.
A custom cell stack model has been integrated in Aspen Plus as a subroutine.
Simulations have been conducted to analyze the performance of an AEL plant.
The proposed model constitutes a useful tool to carry out system optimization.
article info abstract
Article history: A model of an alkaline electrolysis plant is proposed in this paper, including both stack and
Received 16 September 2019 balance of plant, with the objective of analyzing the performance of a complete electrolysis
Received in revised form system. For this purpose, Aspen Plus has been used in this work due to its great potential
13 November 2019 and flexibility. Since this software does not include codes for modelling the electrolysis
Accepted 4 December 2019 cells, a custom model for the stack has been integrated as a subroutine, using a tool called
Available online xxx Aspen Custom Modeler. This stack model is based on semi-empirical equations which
describe the voltage cell, Faraday efficiency and gas purity as a function of the current. The
Keywords: rest of the components in the electrolysis plant have been modelled with standard oper-
Hydrogen production ation units included in Aspen Plus. Simulations have been carried out in order to evaluate
Alkaline water electrolysis and optimize the balance of the plant of an alkaline electrolysis system for hydrogen
Balance of plant production. Also, a parametric study has been conducted. The results show that increasing
System simulation the operation temperature and reducing the pressure can improve the overall performance
Aspen plus model of the system. The proposed model in this work for the alkaline electrolyzer can be used in
Process optimization the future to develop a useful tool to carry out techno-economic studies of alkaline elec-
trolysis systems integrated with other process.
© 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
geno (CNH2), Prolongacio

* Corresponding author. Centro Nacional del Hidro n Fernando El Santo s/n, 13500 Puertollano, Ciudad Real,
Spain.
E-mail addresses: [email protected] (M. Sa nchez), [email protected] (E. Amores), [email protected] (D. Abad),
[email protected] (L. Rodrı́guez), [email protected] (C. Clemente-Jul).
https://doi.org/10.1016/j.ijhydene.2019.12.027
0360-3199/© 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
nchez M et al., Aspen Plus model of an alkaline electrolysis system for hydrogen production, International
Please cite this article as: Sa
Journal of Hydrogen Energy, https://doi.org/10.1016/j.ijhydene.2019.12.027
2 international journal of hydrogen energy xxx (xxxx) xxx
Abbreviations, nomenclature and units r Parameter related to ohmic resistance

(temperature)
Acronyms s Coefficient for overvoltage on electrodes, V
ACM Aspen custom modeler S Entropy, J mol1 C1
AEL Alkaline electrolysis t Coefficient for overvoltage on electrodes
BoP Balance of plant T Temperature, C
CAPEX Capital expenditure V Voltage, V
HTO Hydrogen-to-oxygen W Electric power, W
LHV Low heating value z Electrons transferred per ion
NRTL Non-Random Two-Liquid h Efficiency, %
OTH Oxygen-to-hydrogen b
h Overpotential, V
RES Renewable energy sources
Subscripts
RMS Root mean square
an Anode
Symbols cat Cathode
A Area, m2 cell Alkaline water electrolysis cell
C Parameter related to gas purity (temperature) conc Concentration
d Parameter related to ohmic resistance (pressure) en Energy
E Parameter related to gas purity (pressure) excess Excess or waste
f Parameter related to Faraday efficiency exp Experimental
F Faraday constant, 96485 C mol1 F Faraday
G Gibbs energy, J mol1 gen Generate or produced
H Enthalpy, J mol1 lcoss Loss (via heat radiation)
i Current density, A m2 NET Overall or net
I Current, A ohm Ohmic
n Molar flow rate, mol s1 prod Production
N Number of cells of the stack rev Reversible
p Pressure, bar stack Alkaline water electrolysis stack
Q Thermal power or heat transfer, W th Theoretical or model
tn Thermoneutral
efficiency and durability [2e5]. R&D efforts are carried out in

Introduction order to improve the performance of alkaline electrolysis
technology, mainly focus on:
Renewable energies have had a spectacular development in
recent years. However, a massive penetration of renewable - Development of advanced electrocatalysts to decrease the
energy sources (RES) can affect adversely to the grid stability electrode overvoltage [6e9].
due to its variability and unpredictability [1]. In this context, - Minimization of the space between the electrodes to
hydrogen could play a key role as large-scale energy storage, reduce the ohmic losses and can operate at higher current
through the water electrolysis [2]. densities [10,11].
Currently, alkaline water electrolysis (AEL) is the most - Development of new exchange inorganic membranes to
mature electrolysis technology and it is commercially avail- replace the liquid electrolyte. The use of these membranes
able for large-scale hydrogen production. Typical alkaline allows to reduce the crossover produced trough current
electrolysis cell consists in two Ni-based electrodes immersed porous diaphragms [10,12e14].
in a liquid electrolyte (usually a 30e35 wt% aqueous KOH so-
lution) separated by a porous diaphragm. The operating On the other hand, modelling is an important tool for the
temperature is between 60 and 90 C and the pressure is design and optimization of electrolysis systems [15]. Most of
commonly below 30 bar. The purity of the hydrogen produced alkaline electrolysis models are focused on describing the
is in the range of 99.5e99.9%, which can be increased up to electrochemical behavior of cell/stack. Generally, it consists in
99.999% by catalytic gas purification systems [3]. Therefore, a mathematical description of the polarization curve and
the key to this technology is its availability and the low spe- Faraday efficiency by analytical models [16,17] or empirical
cific cost compared to other electrolysis technologies [4]. models [18e23]. In addition, thermal models are developed
However, the alkaline water electrolysis still presents and integrated with electrochemical models in order to take
certain limitations related to the low current density and the into account the influence of the temperature [16,18,21,23].
influence of the dynamic operation on the gas purity, However, only a few works consider the electrolysis system
international journal of hydrogen energy xxx (xxxx) xxx 3
[20,24e27] and none of the models found in the literature applicable for a wide operating range, a parametric study has
describes the performance of a whole electrolysis plant been conducted, in order to investigate the influence of tem-
including all the components [15]. perature, pressure and current density on the global perfor-
The design and configuration of the balance of plant (BoP) mance of the alkaline electrolysis plant.
and auxiliary systems have a high influence on the perfor- Therefore, this Aspen Plus model could be used as an in-
mance of the electrolysis cells, as well as, cost, efficiency and termediate tool to evaluate alkaline electrolysis plants linked
lifetime. The balance of plant represents approximately 30% to renewable energy sources, since the model is able to predict
of the capital cost (CAPEX) in an alkaline electrolyzer and the the performance of the stack and auxiliary systems at
stack efficiency can be considerably reduced if the design of different loads (power input) in steady operation conditions.
the auxiliary systems is not optimized [4,25,28]. Thus, the In order to evaluate the operation of the complete system
development of models including all components of the plant during the transient periods, a dynamic model could be built
is necessary to optimize alkaline electrolysis systems powered from this using Aspen Dynamics in next steps.
by renewable energy sources, identify technical improve-
ments related to auxiliary systems and reduce its consump-
tion during part-load operation. Aspen Plus model
In this work, a steady state model of an alkaline electrolysis
plant is proposed considering both alkaline water electrolysis The model of an alkaline water electrolysis system has been
cell stack and system. The system includes all the compo- developed using Aspen Plus [35], including alkaline electrol-
nents of the balance of the plant such as deionised water ysis cell stack and the balance of plant (BoP). The simulation
supply, gas-liquid separator vessels, heat exchangers, pumps diagram of the AEL plant studied is shown in Fig. 1.
and the cooling loop. Due to the difficulty of implementing The cell stack (STACK) is the heart of the system. Electricity
mathematical model at a system level in a flexible way, Aspen and heat are supplied to the cells to carry out the decompo-
Plus software has been used in this work. Aspen Plus is one of sition of water into hydrogen and oxygen through an elec-
the most widely used software in the industry for process trochemical reaction shown in Eq. (1):
modelling, equipment design and system optimization. The
1
software is based on a modular operation and multi-flowsheet H2 O / H2 þ O2 (1)
2
architecture enabling the development of easily adaptable
models. Also, it contains libraries of chemicals and pre- Hydrogen (H2-STACK) and oxygen (O2-STACK) produced in
determined unit operation models that simplify greatly the cell stack are led with the electrolyte (KOH, 35%wt) to the
design process [29]. liquid-gas separation vessels (SEP-H2 and SEP-O2, respec-
However, Aspen Plus or Aspen HYSYS do not include codes tively), where the electrolyte is separated from the gas and
for modelling the electrochemical cells, so there are only a few returned back to the stack by recirculation pumps (PUMP-R1
electrolysis models developed using this type of software for cathode circuit and PUMP-R2 for anode circuit). Both KOH
[30e33]. Typically, these works are based on modelling the recycles (ReH2-KOH and ReO2-KOH) pass through a heat
electrolysis cell using standard components available in the exchanger (IC-R1 and IC-R2, respectively) to cool down the
software. However, this approach is not able to describe electrolyte before entering the stack (R-INLET). The cooling
appropriately the electrochemical process. So, novel models circuit is composed by an air-cooler (FAN) and a cooling pump
are required to describe in Aspen Plus the specific electro- (PUMP-COOL) which directs the cooling water (COOL-IN)
chemical phenomena that occur during water electrolysis. through the heat exchangers integrated in the electrolyte
In fact, the novelty of the present study lies here: a semi- recirculation loops to remove the waste heat and maintain the
empirical model for describing the performance of the cell/ temperature in the cells.
stack developed in a previous work [34] has been integrated in The hydrogen and oxygen separated in the biphasic sepa-
Aspen Plus as a subroutine, using a tool called Aspen Custom ration vessels pass through water traps (TRAP-H2 and TRAP-
Modeler (ACM). This custom electrolysis model is able to O2 respectively) to eliminate the maximum amount of
predict in an accurate way, the cell stack voltage, the condensate water. Finally, deionised water with conductivity
hydrogen and oxygen production and the quantity of of 5 mS/cm (H2O-IN) is fed from a water tank into the oxygen
hydrogen in the oxygen (HTO) due to the crossover in function separator (SEP-O2) by a pump (PUMP-H2O) to provide water to
of temperature, pressure and current density. The rest of the electrolysis process (H2O-FEED).
components in the electrolysis plant have been modelled with
standard operation units included in Aspen Plus. Alkaline water electrolysis stack operation unit
So, the integration of this customized model with the rest
of the components of the balance of plant allows having a Since Aspen Plus does not include an operation unit for
complete and innovative model of an alkaline electrolysis modelling an alkaline electrolysis cell stack, in the present
plant. The proposed model constitutes a useful tool to carry study, an alkaline water electrolysis stack model has been
out system level optimization in order to maximize the overall integrated in Aspen Plus as a subroutine, using Aspen Custom
efficiency. Using this mathematical tool, simulations have Modeler (ACM). This tool allows to create a custom operation
been carried out to obtain mass and energy balances of each unit, so the cell stack model is incorporated into the overall
subcomponent and the entire system. The results allow to process to continue with the system simulation [36]. To
analyze the thermodynamic system behavior and optimize develop this, firstly an Aspen Properties file must be created
the overall efficiency. Also, since the model developed is and imported to Aspen Custom Modeler tool in order to define
Fig. 1 e Aspen Plus process flow diagram of an alkaline electrolysis plant.
the components and choose the property method which will Electrochemical model
be used. In this case, a NRTL (Non-Random Two-Liquid) model The electrochemical model for the alkaline electrolysis cells
has been used to carry out the simulations. The different has been developed in a previous work [34]. This model is able
equations, parameters and variables are written in the ACM to predict the electrochemical behavior of an alkaline water
reference language, assuring zero degrees of freedom and electrolysis stack under different operating conditions, such
defining the input/output parameters. as, temperature (T) and pressure (p). The proposed equations
For the stack operation unit developed in Aspen Custom allow determining polarization curve, Faraday efficiency and
Modeler (see Fig. 2), the electrochemical model for the alkaline gas purity, as a function of the current, basing on both phys-
electrolysis cells and all the equations related to the mass and ical principles related to the electrolysis process and statistical
energy balances that take place in the stack have been data [34,37e39].
included. The necessary inputs to carry out simulations will The polarization curve analyzes the different over-
be the electric power input, number of cells, active area of the potentials that occur during the electrolysis of water in order
electrode, stack temperature and operation pressure. to determine the cell potential (Vcell) according to the current
In addition, the ACM tool allows to build a personalized density. For the reaction to occur, a minimum voltage is
icon and connect the different material, heat or work streams. required, which is known as reversible voltage (Vrev), corre-
Once completed this, ACM simulates the model to verify its sponding to 1.23 V at standard conditions (1 bar and 25 C) [37].
performance and export it to the Aspen Plus Library where is However, the cell voltage (Vcell) is always higher than theo-
incorporated into the palette to be used as the rest of the retical one by the appearance of a series of overpotentials due
Aspen Plus standard operation units. to kinetic and resistive effects [23,37,38]. So, the real cell
Fig. 2 e Operation unit developed in Aspen Custom Modeler for modelling the stack.
voltage (Vcell) can be defined as the sum of reversible voltage

Table 1 e Coefficients considered for the electrochemical
and each of these overpotentials (b h ), activation overvoltages
model of an alkaline water electrolysis cell [34].
h cat , b
(b h an ), ohmic overpotentials (b h ohm ) and concentration
Model Coefficient Value Unit
overpotentials (b h conc ) [23,37,38], as shown in Eq. (2).
5
Polarization curve r1 4.45153 10 U m2
Vcell ¼ Vrev þ ðb
h cat þ b
h an þ b
h ohm þ b
h conc Þ (2) r2 6.88874 109 U m2 C1
d1 3.12996 106 U m2
The polarization curve can be determined using a semi-
d2 4.47137 107 U m2 bar1
empirical model. In this regard, one of the most widely used s 0.33824 V
models to describe the electrochemical response of an elec- t1 0.01539 m2 A1
trolyzer was proposed by Ulleberg [18] in 2003. So, taking as t2 2.00181 m2 C A1
reference this model, a more comprehensive equation has been t3 15.24178 m2 C2 A1
developed (Eq. (3)) in a previous work [34], including additional Faraday efficiency f11 478645.74 A2 m4
f12 2953.15 A2 m4 C1
parameters and new empirical relationships to model tem-
f21 1.03960 e
perature (T), pressure (p) and current density (i). Regarding the f22 0.00104 1
C
concentration overpotentials, these occur at very high current Gas purity (hydrogen in C1 0.09901 e
densities, above from the usual range of operation of an elec- 1
oxygen) C2 0.00207 C
trolyzer, so they have not been considered in Eq. (3). C3 1.31064 105 2
C
C4 0.08483 e
1
t2 t3 C5 0.00179 C
Vcell ¼ Vrev þ ½ðr1 þ d1 Þ þ r2 $ T þ d2 $ pi þ s$log t1 þ þ 2 iþ1
T T C6 1.13390 105 2
C
(3) C7 1481.45 A m2
C8 23.60345 A m2 C1
On the other hand, it is possible to measure the effective- C9 0.25774 A m2 C2
ness of the process using the Faraday efficiency (hF), by E1 3.71417 e
comparing the moles produced (nH2, prod) and the theoretical E2 0.93063 bar 1
E3 0.05817 bar 2
moles that should be produced during the same time (nH2, th)
E4 3.72068 e
[38]. This parameter is defined in Eq. (4):
E5 0.93219 bar 1
nH2;prod E6 0.05826 bar 2
hF ¼ (4) E7 18.38215 A m2
nH2;th
E8 5.87316 A m2 bar1
In a similar way to the polarization curve, the Faraday's E9 0.46425 A m2 bar2
efficiency can be also modelled by an empirical expression for
a given temperature using 4 parameters for this purpose such
as Eq. (5) [18,34]. The pressure has not been included due to its The production of oxygen (Eq. (10)) and consumption of
slight influence: water (Eq. (11)) are determined according to the reaction
! stoichiometry (Eq. (1)). Due to its importance, the mass bal-
i2 ance takes into account the amount of hydrogen that is
hF ¼ $ f21 þ f22 $ T (5)
f11 þ f12 $T þ i2 diffused across the diaphragms (Eq. (9)). On the other hand,
since the diffusion of oxygen to hydrogen (OTH) is approxi-
Finally, a model for the diffusion of hydrogen to oxygen
mately 0.1e0.5% [41], has been considered negligible. There-
(HTO) has been proposed in Eq. (6) based on the results ob-
fore, the mass balance for each component is expressed as:
tained in previous works [19,34], considering the influence of
temperature and pressure on the purity of the gases: nH2 ;cat ¼ nH2 ;prod (8)

C7 þC8 $TþC9 $T2
HTO¼ C1 þC2 $TþC3 $T2 þ C4 þC5 $TþC6 $T2 $exp nH2 ;an ¼ nHTO (9)
i

E7 þE8 $pþE9 $p2
þ E1 þE2 $pþE3 $p2 þ E4 þE5 $pþE6 $p2 $exp 1
i nO2 ;an ¼ nO2 ;prod ¼ nH2 ;prod (10)
2
(6)
nH2 O ¼ nH2 ;prod (11)
All coefficients and parameters of the model have been
calculated by means of a non-linear regression using MAT- The gases produced, hydrogen and oxygen, leave the stack
LAB, taking as input experimental data obtained previously immersed in a KOH electrolyte flow, which is latter separated
[23,34]. Table 1 shows the values used in this paper [34]. and recirculated to the stack.
Mass balances Energy balances

The hydrogen production rate at the cathode depends on the The total energy demand for electrolytic hydrogen production
electrochemical behavior of the cells and can be determined is given by the enthalpy of reaction ðDHÞ, according to Eq. (12)
using the Faraday efficiency by Eq. (7) [40]: [3]:
I DH ¼ DG þ TDS (12)
nH2 ;prod ¼ hF $ $N (7)
z$F
When all the energy needed for the electrochemical pro-

cess is provided of electricity, the minimum voltage to carry WNET ¼ Wstack þ WpumpR1 þ WpumpR2 þ WpumpH2O þ WpumpCOOL
out water electrolysis in adiabatic conditions, is the thermo-
þ Wfan
neutral cell voltage (Vtn) [6], according to Eq. (13):
(17)
DH
Vtn ¼ (13) where Wpump-R1, Wpump-R2, Wpump-H2O, Wpump-cool and Wfan de-
z$F
notes power input to the pumps of the system and fan con-
Thus, if cell potential is higher than thermoneutral voltage, sumption, respectively. Wstack is the electric power input for
no external heat is needed for carrying out the electro- the stack operation and can be determined according to Eq
chemical reaction. On the contrary, a heat is generated in the (18), where N is the number of cells of stack:
process directly proportional to the difference between the
cell voltage and the thermoneutral voltage [42], according to Wstack ¼ Vstack $ I ¼ ðVcell $ NÞ$ði $ Acell Þ (18)
Eq. (14). A part of this heat generated is assumed to be lost via
heat radiation from the electrolysis cells (Qloss), thus, the net
excess heat is given by Eq. (15): Results and discussion
Qgen ¼ N $ I$ðVcell Vtn Þ (14) As already discussed, to carry out the simulation in Aspen
Plus, the data required are introduced in the developed model
Qexcess ¼ Qgen Qloss (15) in order to calculate the different outputs. Table 2 lists the
input for the alkaline electrolysis plant simulation at base-
The excess heat results in an increase of the temperature
case conditions.
of electrolyte flow and gases produced in the stack. For this
Also, the following assumptions are considered to simplify
reason, it must be continuously removed in order to ensure a
the simulation process:
constant operating temperature.
All processes operate at steady state

Balance of plant
All the gases in the system behave like ideal gases
10% of the total heat produced by the stack is lost due to
The balance of plant includes all the equipment needed to
convection and radiation
operate the stack, such as, deionised water supply, heat ex-
Liquid deionised H2O is fed to the system in a reference
changers, gas-liquid separator vessels, circulation pumps and
environment condition at 298 K
the cooling loop. As it can be seen in Fig. 1, the process flow
The hydrogen and oxygen output are at 298 K
model has been developed to include all the major compo-
The electrolyte recirculation is fixed at 15 l min1
nents that would be present in an actual alkaline electrolysis
AEL stack is operated at balanced anode and cathode
plant.
pressure
Most components in the balance of the plant have been
modelled using the standard elements found in Aspen Plus
software. Introducing system boundary conditions, individual
Experimental model validation
component efficiencies and operating parameters to the pro-
cess, the Aspen software performs energy and mass balances
For carrying out the experimental tests and the validation of
across all components to predict fluid conditions around the
the model, a cell stack composed of 12 bipolar alkaline elec-
system including thermodynamic data for all chemical spe-
trolysis cells of 1000 cm2 surface area connected electrically in
cies involved in the process [35].
series, has been characterized at different operating condi-
tions into a fully automated test bench developed by Centro
System efficiency
Nacional del Hidrogeno (CNH2).
This test bench (Fig. 3), which has been described in
To assess the overall performance of the system, the energy
detailed in previously published papers [34,44e46], has a ca-
efficiency should be calculated. The energy efficiency rep-
pacity of 2.5 Nm3 h1 of hydrogen production and has been
resents the ratio of the energy contained in the useful
designed to be able to operate in a wide range of temperatures
products of a process to the energy contained in all input
streams [43]. Therefore, the energy efficiency for hydrogen
production from an electrolysis system is formulated as
follows (Eq. (16)) [42]: Table 2 e Input data for the system simulation at base-
case operation conditions (75 C and 7 bar).
nH2;OUT ,LHVH2 Parameter Value Unit
hen ¼ (16)
WNET

Stack working temperature, Tstack 75 C
where LHVH2 is the hydrogen lower heating value; nH2, OUT is Operating system pressure, pstack 7 bar
the outlet flow rate of hydrogen in the electrolysis plant; WNET Electrolyte concentration 35 wt% KOH
is the electric power input to the AEL system. For the complete Active electrode area, Acell 1000 cm2
alkaline electrolysis system shown in Fig. 2, the net power is Cell number, N 12 cells
Input power stack, Wstack 10 kW
defined by Eq. (17):
Fig. 3 e 15 kW alkaline water electrolysis test-bench developed by CNH2 [46].
(40e80 C) and pressures (1e10 bar). Two programmable DC The experimental data obtained in this test bench (po-
switching power supplies provide electrical power to the cells, tential, hydrogen production and gas purity) have been
with the possibility of being connected in series or parallel, used to calculate the coefficients shown in Table 1, as
reaching a maximum power of 15 kW (0e120 V and 0e500 A). previously mentioned [34]. Subsequently, simulated and
The purity of the gases produced is monitored by on-line gas measured values have been compared to ensure the accu-
analyzers. An accurate thermal mass flow meter has been racy and validity of the proposed model using the root-
installed at the hydrogen output for measuring hydrogen mean-square (RMS) error. Fig. 4 shows a parity chart of
production. Also, water flow rate of the cooling system for the the model and experimental results used in this paper. The
electrolyzer is logged and collected. For thermal character- results show an excellent correlation between experimental
ization, thermocouples are included in all the main streams of and modelled data: the RMS error is approximately 5 mV
the system. The plant control unit is based on a PLC (pro- per cell and lower than 1% for the Faraday efficiency and
grammable logic controller) that controls the system through HTO [34]. This demonstrates the usefulness of the model to
a series of strategies optimized and acquires all the informa- predict the electrolyzer response in other experimental
tion provided by the different sensors. conditions.
Fig. 4 e Parity chart of the model and experimental results: a) Stack voltage; b) Faraday efficiency; c) Gas purity.
Alkaline water electrolysis cell stack performance In Fig. 5b, the ohmic and activation overpotentials are
showed individually. The ohmic overpotentials are related
In this section, a study has been conducted using the Aspen with the electrical resistance of the different materials of the
Plus model developed in this work to investigate the effect of electrolysis cell and the interfaces between them; while the
operating temperature, pressure and current density on the activation overpotentials are as a consequence of the activa-
performance of the cell stack. In the figures, all other param- tion energy of reactions that occur on the electrodes. When
eters are fixed at base-case conditions (see Table 2). With this the current density increases both overpotentials grow
objective, it has been evaluated the polarization curve, over- significantly, so that the overall potential is increased. As it
potentials in the cell, power required by the stack, heat can be seen the activation overpotentials dominate the
generated and required and the hydrogen crossover (HTO), voltage losses [47]. When the temperature increases, the
among others. activation overpotentials are reduced because the electro-
chemical reaction is faster [47]. On the other hand, the high
Influence of the temperature and current density conductivity of the electrolyte (KOH 35%) and stack design
Fig. 5a shows the polarization curve and power required by the lead to a lower ohmic overpotential [48].
alkaline electrolysis stack at different temperatures. Accord- In order to analyze the heat requirements for the AEL stack
ing to the model, when the temperature increases from 50 C operation is important to evaluate the heat production by the
to 80 C, the voltage progressively reduces. As a consequence, irreversibilities. As it can be seen in Fig. 5c, the heat produc-
the stack power required in the electrolysis decreases when tion (due mainly to activation overpotentials) increases sig-
the temperature is higher. nificatively with the current density and is lower at higher
Fig. 5 e Effect of temperature on an AEL stack performance at 7 bar: a) Polarization curve and stack power required; b) Cell voltage
and overpotentials; c) Heat generated, heat required and excess heat; d) Hydrogen flow rate and hydrogen crossover (HTO).
temperature. For a better understanding, the theoretical On the one hand, when the pressure raises the void frac-
thermal energy demand ðTDSÞ is also shown in the figure for tion between electrodes is reduced due to smaller size of
comparison. It is observed that, in all the range studied, the generated gas bubbles. This reduction in bubbles sizes also
heat production by overpotentials exceeds the energy affects the effective contact area between electrodes and
required by the electrolysis process in adiabatic conditions. It electrolyte, which results in lower ohmic resistance and
means that, in this case, no external heat input is needed and required cell voltage [51]. On the other hand, the reversible cell
the excess heat must be cooled in order to maintain the voltage increases with the pressure. As a consequence, in the
operation temperature constant. pressure range studied, these phenomena are counter-
Fig. 5d shows the hydrogen production and the content of balanced and ohmic overpotentials increase slightly when the
hydrogen in oxygen (HTO) at different temperatures. The pressure increases. Regarding the activation overpotentials,
hydrogen production increases when the temperature is the pressure has practically no perceptible influence.
reduced. This is because an increase in temperature leads to a In Fig. 6c is illustrated the heat generated, heat required
lower resistance, more parasitic current losses and so lower and excess heat at different pressures. As shown, the heat
Faraday efficiencies [18]. Regarding to the content of hydrogen production by irreversibilities increases at low pressures
in oxygen (HTO), the results show that high temperatures have because although the stack voltage increases slightly from 5 to
a great influence on the generated impurities, since the 9 bar, as discussed previously, the theoretical thermal energy
diffusion phenomena and the gas migrations increase signif- demand ðTDSÞ is reduced at higher pressures. As with tem-
icantly. On the other hand, as can be seen in Fig. 5d, at low perature (Fig. 5c), the heat production by overpotentials ex-
current densities, the purity of the gases produced is signifi- ceeds the energy required by the electrolysis process in
cantly reduced. This is due to that the mentioned phenomena adiabatic conditions for any pressure. As consequence, the
are mostly independent of the electrolyzer load, so when the excess heat must be removed to maintain the temperature in
production gas rate is low the percentage of total impurities is the cells.
higher [34,39]. Thus, for safety reasons, the minimum applied Regarding to the content of hydrogen in oxygen (HTO), as
current density in the tests was 0.1 A cm2. can be seen in Fig. 6d, it increases strongly with decreasing the
current density for all pressures. In addition, the results show
Influence of the pressure and current density that a higher pressure increases the impurities. This occurs
Fig. 6a shows the polarization curve and power required by the because the pressure has a strong influence both on the sol-
alkaline electrolysis stack at different pressures. The main ubility of the gases in the electrolyte and on the gas diffusion
variation occurs in the ohmic overpotentials (Fig. 6b) due to a between the cathode and the anode inside the cell. About
series of phenomena mainly related to the size of the gas hydrogen flow rate, the effect of the pressure can be consid-
bubbles generated during the electrolysis [34,49,50]. ered negligible.
Fig. 6 e Effect of pressure on an AEL stack performance at 75 C: a) Polarization curve and stack power required; b) Cell voltage
and overpotentials; c) Heat generated, heat required and excess heat; d) Hydrogen flow rate and hydrogen crossover (HTO).
Overall system analysis It can be observed from Fig. 8a that, with the increase of
current density the difference between the stack power and
In this section, energy and mass balances of each component the net system power increases, which is due to the increase
and the entire system are calculated using Aspen Plus simu- in parasitic loads. As a consequence, Fig. 8b shows that the
lation. Fig. 7 shows the flow diagram of the alkaline electrol- consumption of the balance of the plant reduces the stack
ysis plant with the results of the simulation at base case efficiency considerably and the reduction is higher when
operation conditions (Table 2). Mass balances and enthalpies current density increases. At the specified operation point of
at specified conditions are shown, also heat losses and elec- 0.42 A cm2, stack efficiency is 57.9% and system efficiency is
trical consumption of different component are included. For 53.3%. Also, in Fig. 8b can be seen that the efficiency initially
easy understanding, the values of pressure and temperature increases at lower current densities reaching a peak and
of each stream are represented. The detailed composition of finally decreases with the increase of the current density.
each matter flow is shown in Table 3. There is an operation point where system efficiency is maxi-
From data obtained, it can be seen that when the system mized and the specific consumption reaches the minimum
operates at current density of 0.42 A cm2 (corresponding to a point. The maximum system efficiency obtained is 55.5% at a
power required by the electrolysis process of 10 kW), it leads to current density of 0.25 A/cm2.
a hydrogen production rate of 1.95 Nm3 h1 (0.17 kg h1, In Fig. 8c, the distribution of the electric power input is
without taking into account the water in the stream) at 25 C indicated at 0.42 A cm2 and an analysis of the consumption
and 1 atm. As for the energy balance, the results show that the of the auxiliary systems (pumps and fan) at the same point is
energy is perfectly balanced in the stack and the waste heat is shown in Fig. 8d, considering different temperatures and
indirectly retired in heat exchangers located in the electrolyte pressures. As it can be observed, the power required by the
recirculation loops. The total excess heat is calculated to be balance of the plant is less than 1 kW, increasing at lower
2392.8 W and it is dissipated by cooling water from an air- temperature and higher pressure.
cooler system at 35 C. Finally, Fig. 9 shows the variation of the system efficiency
In order to analyse how the consumption of auxiliary with the temperature and pressure in an operating map at
components impacts on the efficiency system, the variation of current density of 0.25 and 0.42 A cm2 (corresponding to an
net system power and stack power with current density at electrolysis input power of 6 kW and 10 kW). As can be seen,
base-case conditions is presented in Fig. 8a. when the temperature increases and the pressure diminishes,
Fig. 7 e Energy and mass flow diagram of the system at base case operation conditions (10 kW, 75 C and 7 bar).
Table 3 e Composition of the matter flow at base-case operation conditions (75 C and 7 bar).
Stream T P Mass flow (kg h1) Composition (kg h1)
( C) (bar)
H2O H2 O2
H2O-IN 25 1 1.730 1.730 0 0
H2O-FEED 25.49 7 1.730 1.730 0 0
H2-STACK 75 7 447.650 447.470 0.178 0
O2-STACK 75 7 448.870 447.470 0.000131 1.404
ReO2-KOH 74.81 6.7 449.230 449.150 0 0.07506
ReH2-KOH 74.99 6.7 447.380 447.370 0.00408 0
R-INLET 72.75 7 896.610 896.530 0.00409 0.07507
H2-PROD 74.99 6.7 0.260 0.095 0.173 0
H2-OUT 25 1 0.220 0.050 0.173 0
PURG-1 25 1 0.044 0.044 0 0
O2-PROD 74.81 6.7 1.374 0.045 0.000125 1.329
O2-OUT 25 1 1.353 0.024 0.000125 1.329
PURG-2 25 1 0.021 0.021 0 0
Fig. 8 e Energy system analysis at base-case conditions: (a) stack power, net power and auxiliary power; (b) stack and
system energy efficiencies; (c) distribution of electric power input by components; (d) analysis of auxiliary system
consumption: influence of the pressure and temperature.
Fig. 9 e Operating map of the alkaline electrolysis system efficiency at an electrolysis input power of 6 kW and 10 kW.
the overall efficiency improves. In this way, for example, if the stack efficiency up to 5%, being the reduction higher when
operating temperature is 60 C, a similar efficiency is obtained current density increases. In particular, at reference condi-
using 5 bar at 0.42 A cm2 that 9 bar at 0.26 A cm2. On the tions (7 bar and 75 C) and current density of 0.42 A cm2, stack
other hand, when the pressure is 8 bar the efficiency is efficiency is approximately 58% and system efficiency is
approximately the same at 75 C and 0.42 A cm2 than at 60 C 53.3%.
and 0.26 A cm2. Finally, an operating map has been used to quantify the
Therefore, the efficiency can be kept constant by most influential parameters in the overall system efficiency. It
decreasing the pressure if the current density is increased identifies which variable should be modified to achieve a
(always above operation point where system efficiency is greater efficiency. In general terms, the results indicate that
maximized), or if the temperature and current density in- the influence of the pressure is not as strong as the operation
crease at the same time. So, this map establishes which temperature in the investigated ranges. So, when the tem-
parameter would have to be changed to reach a greater effi- perature increases and the pressure diminishes, the overall
ciency and it is a first indicator of process performance. efficiency improves and the energy consumption is reduced.
Particularly, for this alkaline electrolysis system the optimum
operation conditions would be 5 bar and 80 C reaching an
Conclusions overall system efficiency close to 58% at 0.25 A/cm2.
In conclusion, the novel Aspen Plus proposed in this
In this paper, an Aspen Plus model of an alkaline water elec- work provides a useful design tool that can be applied to
trolysis plant has been proposed with the objective of evalu- maximize the efficiency and cost of an alkaline electrolysis
ating the performance of a complete system, including the system powered by renewable energy sources for hydrogen
stack and balance of plant, under different operating condi- production. The obtained results allow determining which
tions, such as temperature and pressure. process variables are the most influential and hence,
For this purpose, a custom model of the electrolysis cells should be optimized in order to improve the performance
developed in a previous work has been integrated in Aspen and operation. This model can further be used in the future
Plus as a subroutine using Aspen Custom Modeler. This stack to develop a powerful tool to perform techno-economic
model is based on semi-empirical equations that describe the analysis of alkaline electrolysis systems integrated with
voltage cell, Faraday efficiency and gas purity as a function of other process.
the current density. The rest of the components in the elec-
trolysis plant have been modelled with standard operation
units included in Aspen Plus.
In order to ensure the accuracy and validity of the model, a Acknowledgements
comparison between the experimental data and the calcu-
lated values by the proposed model has been carried out. The The work described in this paper has been developed in the
RMS error is approximately 5 mV per cell and lower than 1% geno (CNH2), whose
facilities of the Centro Nacional del Hidro
for the Faraday efficiency and HTO. So, the results show an financial supporters are Ministerio de Ciencia, Innovacio n y
excellent correlation, which demonstrates the usefulness of Universidades (MCIU, Spain), Junta de Comunidades de
the model to predict the electrolyzer response in other Castilla-La Mancha (JCCM, Spain) and European Regional
experimental conditions. Development Fund (ERDF).
On the basis of this, a parametric study has been con-
ducted. It is observed that when the temperature increases
from 50 C to 80 C, the voltage progressively reduces. As a references
consequence, the stack power required in the electrolysis
decreases. On the other hand, an increase in the temperature
leads to lower hydrogen rate due to that the Faraday efficiency [1] Kotowicz J, Jurczyk M, Wecel D, Ogulewicz W. Analysis of
is reduced and higher crossover of hydrogen to the oxygen hydrogen production in alkaline electrolyzers. J Power
Technol 2016;96(3):149e56.
side (HTO). With regard to the pressure, the stack voltage in-
[2] Lehner M, Tichler R, Steinmüller H, Koppe M. Power-to-Gas:
creases slightly when the pressure increases in the range technology and business model. Cham (Switzerland):
studied (from 5 to 9 bar) and no variations are observed in Springer; 2014.
hydrogen production. Also, the higher the pressure the higher [3] Buttler A, Spliethoff H. Current status of water electrolysis
the impurity of the gases produced. for energy storage, grid balancing and sector coupling via
In addition, the Aspen Plus model presented in this work power-to-gas and power-to-liquid: a review. Renew Sustain
Energy Rev 2018;82:2440e54.
enables to obtain mass and energy balances of each sub-
[4] Bertuccioli L, Chan A, Hart D, Lehner F, Madden B, Standen E.
component and the entire system in a fast and accurate way.
Development of water electrolysis in the European Union.
On the basis of this, it has been possible to analyse the ther- Cambridge (UK): FCH JU funded studies; 2014. Element
modynamic system behaviour and identify the consumption Energy, Lausanne (CH); E4tech Sarl.
of each component as well as energy losses sources. It has [5] Ursua A, Gandı́a LM, Sanchis P. Hydrogen production from
been observed that the power required by the balance of the water electrolysis: current status and future trends. Proc IEEE
plant is approximately 8% of the net power and reduces the 2012;100(2):410e26.
[6] Michal J, Daniel W, Zeng K, Zhang D. Recent progress in alkaline water electrolyzer: experimental study and
alkaline water electrolysis for hydrogen production and mathematical modelling. Int J Hydrogen Energy
applications. Prog Energy Combust Sci 2010;36(3):307e26. 2008;33(24):7338e54.
[7] Li X, Walsh FC, Pletcher D. Nickel based electrocatalysts for [27] Zhang H, Su S, Lin G, Chen J. Configuration design and
oxygen evolution in high current density alkaline water parametric optimum criteria of an alkaline water
electrolysers. Phys Chem Chem Phys 2011;13:1162e7. electrolyzer system for hydrogen production. Int J
[8] Singh RN, Mishra D, Sinha ASK, Singh A. Novel Electrochem Sci 2011;6:2566e80.
electrocatalysts for generating oxygen from alkaline water [28] Schmidt O, Gambhir A, Staffel I, Hawkes A, Nelson J,
electrolysis. Electrochem Commun 2007;9:1369e73. Few S. Future cost and performance of water electrolysis:
[9] Pletcher D, Li X, Wang S. A comparision of cathodes for zero an expert elicitation study. Int J Hydrogen Energy
gap alkaline water electrolyzers for hydrogen production. Int 2017;42:30470e92.
J Hydrogen Energy 2012;37:7429e35. [29] Anderson T, Vijay P, Tade MO. An adaptable steady state
[10] Li X, Pletcher D. Prospects for alkaline zero gap water Aspen Hysys model for methane fueled oxide fuel cell. Chem
electrolysers for hydrogen production. Int J Hydrogen Energy Eng Res Des 2014;92:295e307.
2011;36:15089e104. [30] Hauck M, Herrmann S, Spliethoff H. Simulation of a
[11] Manabe A, Kashiwase M, Hashimoto T, Hayashida T, Kato A, reversible SOFC with aspen Plus. Int J Hydrogen Energy
Hirao K, Shimomura I, Nagashima I. Basic study of alkaline 2017;43:10329e40.
water electrolysis. Electrochim Acta 2013;100:249e56. [31] Rivera-Tinoco R, Farran M, Bouallou C, Auprete F, Valentin S,
[12] Leng Y, Chen G, Mendoza AJ, Tighe TB, Hickner MA, Millet P, Ngameri JR. Investigation of power-to-methanol
Wang MA. Solid-state water electrolysis with an alkaline processes coupling electrolytic hydrogen production and
membrane. J Am Chem Soc 2012;134:9054e7. catalytic CO2 reduction. Int J Hydrogen Energy
[13] Xiao L, Zhang S, Pan J, Yang C, He M, Zhuang L, Lu J. First 2016;43:4546e59.
implementation of alkaline polymer electrolyte water [32] JaeHwa K, DuckJoo Y, Chang HO. Simple electrolyzer model
electrolysis working only with pure water. Energy Environ development for high-temperature electrolysis system
Sci 2012;5:7869e71. analysis using solid oxide electrolysis cell. J Nucl Sci Technol
[14] Pavel CC, Cecconi F, Emiliani C, Santiccioli S, Scaffidi A, 2010;47(7):590e607.
Catanorchi S, Comotti M. Highly efficient platinum group [33] Stoots CM, James EO, McKellar MG, Hawkes GL. Engineering
metal free based membrane-electrode assembly for anion process model for high-temperature electrolysis system
exchange membrane water electrolysis. Angew Chem Int Ed performance evaluation. AIChe 2005 Annual meeting. Idaho
2014;53(5):378e81. National Laboratory; 2005.
[15] Olivier P, Bourasseau C, Bouamama PB. Low-temperature [34] Sa nchez M, Amores E, Rodrı́guez L, Clemente-Jul C. Semi-
electrolysis system modelling: a review. Renew Sustain empirical model and experimental validation for the
Energy Rev 2017;78:280e300. performance evaluation of a 15 kW alkaline water
[16] Hammoudi M, Henao C, Agbossou K, Dube Y, Doumbia ML. New electrolyzer. Int J Hydrogen Energy 2018;43(45):20332e45.
multi-physics approach for modelling and design of alkaline [35] Hanyack ME. Chemical process simulation and the Aspen
electrolyzers. Int J Hydrogen Energy 2012;37:13895e913. HYSYS v8.3. Software. CreateSpace Independent Publishing
[17] Heano C, Agbossou K, Hammoudi M, Dube Y, Cardenas A. Platform; 2013.
Simulation tool based on a physics model and an electrical [36] Dyment J, Mantrala V. Jump start: Aspen custom modeler
analogy for an alkaline electrolyser. J Power Sources v8.0. A brief tutorial. Aspen Technology, Inc.; 2015.
2014;250:58e67. [37] Gandı́a LM, Arzamendi G, Die guez PM. Renewable
[18] Ulleberg Ø. Modeling of advanced alkaline electrolyzers: a hydrogen technologies. Production, purification, storage,
system simulation approach. Int J Hydrogen Energy applications and safety. Amsterdam: Elsevier Publishing
2003;28(1):21e33. Company; 2013.
[19] Hug W, Bussmann H, Brinner A. Intermittent operation and [38] Amores E, Rodrı́guez J, Oviedo J, de Lucas-Consuegra A.
operation modeling of an alkaline electrolyzer. Int J Development of an operation strategy for hydrogen
Hydrogen Energy 1993;18(12):973e7. production using solar PV energy based on fluid dynamic
[20] Busquet S, Hubert CE, Labbe J, Mayer D, Metkemeijer R. A aspects. Open Eng 2017;7:141e52.
new approach to empirical electrical modelling of a fuel cell, [39] Haug P, Koj M, Turek T. Influence of process conditions on
an electrolyser or a regenerative fuel cell. J Power Sources gas purity in alkaline water electrolysis. Int J Hydrogen
2004;13:41e8. Energy 2017;42(15):9406e18.
[21] Vanhanen JP, Lund PD. Computational approaches for [40] Ursua A, Marroyo L, Gubı́a E, Gandı́a LM, Die guez PM,
improving seasonal storage systems based on hydrogen Sanchis P. Influence or the power supply on the energy
technologies. Int J Hydrogen Energy 1995;20(7):575e85. efficiency of an alkaline water electrolyser. Int J Hydrogen
[22] Artuso P, Gammon R, Orecchini F, Watson SJ. Alkaline Energy 2009;34:3221e33.
electrolysers: model and real data analysis. Int J Hydrogen [41] Janssen H, Bringmann JC, Emonts B, Schroeder V. Safety-
Energy 2011;36(13):7956e62. related studies on hydrogen production in high-pressure
[23] Amores E, Rodrı́guez J, Carreras C. Influence of operation electrolysers. Int J Hydrogen Energy 2004;29:759e70.
parameters in the modeling of alkaline water electrolyzers [42] Ni M, Leung MKH, Leung DYC. Energy and exergy analysis of
for hydrogen production. Int J Hydrogen Energy production by solid oxide steam electrolyzer plant. Int J
2014;39(25):13063e78. Hydrogen Energy 2007;32:4648e60.
[24] Zhou T, Francois B. Modeling and control design of hydrogen [43] Al-Ghandoor A, Phelan PE, Villalobos R, Jaber JO. Energy and
production process for an active hydrogen/wind hybrid exergy utilizations of the U.S. manufacturing sector. Energy
power system. Int J Hydrogen Energy 2009;3:21e30. 2010;35:3048e65.
[25] Mori M, Mrz ljak T, Drobnic B, Sekavc
nik M. Integral [44] Sa nchez M, Amores E, Rodrı́guez J, Sa nchez-Molina M.
characteristics of hydrogen production in alkaline Laboratorio de Electro lisis Alcalina del Centro Nacional del
electrolysers. J Mech Eng 2013;59(10):585e94. Hidro geno: una instalacio n al servicio de la comunidad
[26] Dieguez PM, Ursúa A, Sanchis P, Sopena C, Guelbenzu E, gica e industrial. In: Libro de
cientı́fico-tecnolo
Gandı́a LM. Thermal performance of a commercial Comunicaciones del Congreso Iberoamericano de Hidro geno
y Pilas de Combustible, Barcelona (Spain); October 15-17; [48] Tijani AS, Yusup NAB, Rahim AHA. Mathematical modelling
2014. and simulation analysis of advanced alkaline electrolyzer
nchez M, Hidalgo D, Quintana R, Saiz F. Influence of the
[45] Sa system for hydrogen production. Procedia Technology
design of liquid-gas separators in the operation of an 2014;15:799e807.
alkaline water electrolysis system for hydrogen production: [49] Ursúa A, Sanchis P. Static-dynamic modelling of the
evaluation and optimization. In: 2nd ADEL International electrical behaviour of a commercial advanced alkaline
Workshop, Ajaccio (France); May 8e10; 2013. water electrolyser. Int J Hydrogen Energy 2012;37:18598e614.
nchez M. Desarrollo de sistemas de produccio
[46] Sa n de [50] Onda K, Kyakuno T, Hattori K, Ito K. Prediction of production
hidro geno energe tico por generacio n alcalina: proyecto power for high-pressure hydrogen by high-pressure water
DESPHEGA. Revista Energe tica XXI 2013;136:26e7. electrolysis. J Power Sources 2004;132:64e70.
[47] Milewski J, Guandalini G, Campanari S. Modeling an alkaline [51] Mazloomi SK, Sulaiman N. Influencing factors of water
electrolysis cell through reduced-order and loss estimate electrolysis electrical efficiency. Renew Sustain Energy Rev
approaches. J Power Sources 2014;269:203e11. 2012;16:4257e63.

Aspen Plus Model of An Alkaline Electrolysis System For Hydrogen Production

Uploaded by

Copyright:

Available Formats

Aspen Plus Model of An Alkaline Electrolysis System For Hydrogen Production

Uploaded by

Document Information

Original Title

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

Aspen Plus Model of An Alkaline Electrolysis System For Hydrogen Production

Uploaded by

Copyright:

Available Formats

international journal of hydrogen energy xxx (xxxx) xxx

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he