Energy Conversion and Management 258 (2022) 115499

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Techno-economic analysis and Monte Carlo simulation of green hydrogen

production technology through various water electrolysis technologies
Dohyung Jang a, Jaedong Kim b, Dongmin Kim b, Won-Bi Han b, Sanggyu Kang a, *
a
Department of Naval Architecture and Ocean Engineering, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Republic of Korea
b
H2 Technology R&D Division, KOGAS Research Institute, Korea Gas Corporation, 1248 Suin-ro, Sangnok-gu, Ansan 15328, Republic of Korea

A R T I C L E I N F O A B S T R A C T

Keywords: The green hydrogen produced by water electrolysis combined with renewable energy is inevitable for a sus­
Hydrogen production tainable and clean energy society. However, the green hydrogen production cost is still more expensive than
Water electrolysis fossil fuel-based methods such as steam methane reforming and it is the biggest challenge. In this study, to
Economic evaluation
determine the most economical way to obtain green hydrogen, a technoeconomic analysis was performed on four
Levelized cost of hydrogen
Sensitivity analysis
different water electrolysis technologies, that is, alkaline water electrolysis (AWE), proton exchange membrane
Monte Carlo simulation method electrolysis (PEMEC), solid oxide electrolysis with electric heaters (SOEC (E.H)), and solid oxide electrolysis
combined with a waste heat source (SOEC (W.H)). It is assumed that the SOEC (W.H) system is linked to the
adjacent power plant and receives waste heat. The net present value calculation, sensitivity analysis, and Monte
Carlo method were performed for the 4 cases presented above. Calculated unit hydrogen production cost was
7.60, 8.55, 10.16, and 7.16 $/kgH2 for AWE, PEMEC, SOEC (E.H), and SOEC (W.H), respectively. The analysis
results show that the SOEC (W.H) has the most competitive case due to sensible heat energy saving and higher
stack efficiency. Sensitivity analysis determined the most influential cost factor, and consequently, it seems
necessary to reduce capital costs along with subsidizing electricity costs and tax rates to compete with gray
hydrogen.

especially the power-to-gas (P2G) method, is known as one of the most

1. Introduction promising solutions to store excess energy in RESs. Therefore, hydrogen,
which is capable of large-scale storage and is advantageous for long-
It is necessary to transition from the current fossil-based energy so­ term storage without severe losses [3], is attracting attention as a key
ciety to a new green energy society to suppress greenhouse gas emissions technology for storing clean and sustainable energy.
while responding to the increasing energy demand. Renewable energy Hydrogen is sustainable, highly abundant, and non-toxic. It has the
plays an important role in addressing this problem. Therefore, the pol­ highest energy content per weight (for lower heating value, 120 MJ/kg
icies of the world’s major developed countries aim to improve the pro­ [4]) among any fuel and releases only water as a by-product when it is
portion of energy generated by renewable energy sources (RES). As a burned. Therefore, it is regarded as a clean, renewable fuel and,
result of such efforts, in Europe, wind and solar power generation consequently, it can reduce the dependence on oil reserves in the Middle
exceeded the amount of coal power generation in 2019, and the gen­ East. However, hydrogen also is not an impeccable method. One of the
eration by total RES exceeded that of fossil fuel in 2020, accounting for demerits is low volumetric energy density (4 times less than gasoline
40% and 34%, respectively [1]. [4]), which can be an obstacle to storage. To overcome storage limita­
At this point, the main problem to solve is that power generation tions, metal hydrides, gas hydrates, physisorption have been studied.
from RES is vastly weather-dependent. This instability leads to fluctu­ More recently, offshore geological storage has also been proposed [5–7].
ations in energy generation and imbalances in supply and demand. As a High flammability is another problem. Hydrogen is ignited much easier
result, energy storage strategies have become increasingly important, than most other fuels with a wide range of flammability (4–75 vol%) [8].
and research on various energy storage methods, such as pumped hy­ However, hydrogen does not easily ignite unless it is in an enclosed,
dropower, secondary batteries, compressed air energy storage, and unventilated space due to its nature to disperse quickly into the air [9].
power-to-gas storage, is actively being conducted [2]. Hydrogen storage, Moreover, due to superior hydrogen detection sensors, hydrogen has

* Corresponding author.
E-mail address: [email protected] (S. Kang).

https://doi.org/10.1016/j.enconman.2022.115499
Received 25 November 2021; Received in revised form 9 February 2022; Accepted 12 March 2022
Available online 21 March 2022
0196-8904/© 2022 Elsevier Ltd. All rights reserved.
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

Nomenclature n Operating year of plant

n Cell number
A Active area of electrolyzer cell [cm2] NPV Net present value
AWE Alkaline water electrolysis OPEXn Operating expenditure at operating year n
BOP Balance of Plants PH2 Selling price of hydrogen
CCS Carbon capture and storage PO2 Selling price of oxygen
CFn Net cash flow at operating year n PEMEC Proton exchange membrane electrolysis
F Faraday’s constant [96,485 C/mol] r Discount rate
HHV Higher heating value [39.4 kWh/kg] RES Renewable energy source
LHV Lower heating value [33.3 kWh/kg] REV n Revenue from the products sales at operating year n
i Current density [A/cm2] REVHn Revenue from the hydrogen sales at operating year n
I0 Capital expenditure (CAPEX) REVOn Revenue from the oxygen sales at operating year n
IRR Internal ratio of return SMR Steam methane reforming
k An indicator whether oxygen is sold or not SOEC (W.H) Solid oxide electrolysis coupled with waste heat
LCOE Levelized cost of energy sources
LCOH Levelized cost of hydrogen SOEC (E.H) Solid oxide electrolysis with electric resistance heaters
M Hydrogen production rate [mol/sec] SRD System degradation rate
MH2 Annual production of hydrogen in mass [kg/year] TR Tax rate
MCM Monte-Carlo method

Fig. 1. Operating principles of alkaline water electrolysis, proton exchange membrane electrolysis, and solid oxide electrolysis.

been utilized safely in the industry for a long time [10]. Meanwhile, Current hydrogen production is primarily used by steam methane
when more hydrogen is emitted and its concentration in the atmosphere reforming (SMR) of natural gas (48%) or utilization of other fossil fuels
increases, hydrogen reacts with hydroxyl (OH) radicals as in Eq. (1). (48%) These fossil-based hydrogens, which are called gray hydrogen,
are not our long-term goal since they still lead to 500 million tons of
H2 + OH→H2 O + H (1)
carbon dioxide emissions annually. In addition to blue hydrogen, which
This reaction could have the following two serious consequences has significantly reduced greenhouse gas emissions by applying carbon
[11–14]. First, the water vapor produced by the reaction depletes the capture and storage (CCS) technology [19], green hydrogen produced by
stratospheric ozone layer through its moistening. Second, the reaction water electrolysis coupled with renewable energy sources, which
reduces the concentration of hydroxyl (OH) radicals which also reacts currently accounts for only 4% of hydrogen production, is expected to
with and removes other chemicals in the atmosphere. One of the pre­ increase its proportion gradually and account for 22% of total hydrogen
sentative chemicals is methane which is a potent greenhouse gas. production by 2050 [20].
However, recent studies anticipate that the global warming impact by
emitted hydrogen will not be severe [15–17]. Despite the above po­ - Fig. 1 shows the operating principles of three water electrolysis
tential problems, the huge advantages of hydrogen are so attractive that technologies. The overall reaction is identical while there is a dif­
many national governments are making effort to expand hydrogen ference in the type of moving cation or anion. These three technol­
application and build a hydrogen society [18]. ogies are mainly being developed, and the characteristics of each
technology are as follows [21]: Alkaline water electrolysis (AWE):

2
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

AWE has been studied for the longest time Thus, it has the highest AWE, PEMEC, and SOEC. The SOEC has been considered a being
level of technical maturity and is already commercially available. developed at the laboratory level, thus only AWE and PEMEC have been
AWE uses KOH or NaOH aqueous solution as an electrolyte to compared for practical analysis. However, research has continued to
conduct a water decomposition process in an alkaline environment. improve the challenges, and Sunfire has presented a 750 Nm3/hr class
It typically operates at 60–80 ◦ C under a pressure of 30 bar. It has SOEC system. Therefore, now SOEC also needs to be compared together.
high durability and does not require novel catalysts, so the produc­ Ferrero et al. [30] investigated three water electrolysis systems for the
tion cost is low. Therefore, price competitiveness is an advantage in power-to-gas option and presented the result with the levelized cost of
planning large-scale electrolysis plants. However, it uses a relatively electricity (LCOE) of stored energy. Tenhumberg et al. [31] presented an
permeable separator that can allow crossover of the generated bub­ evaluation of three water electrolysis systems in terms of the carbon
bles. Thus, it has a low hydrogen purity and a limit to pressurized footprint. Both studies evaluated AWE systems as the most cost-effective
operation. Above all, the biggest problem is that it is not flexible technology, presenting the results of the calculation of hydrogen pro­
enough to accept renewable energy sources due to its slow response duction costs to help compare each technology’s economic feasibility.
to intermittent input. Therefore, the development objective is However, the former study assumed the electric cost to be free by uti­
focused on system design for dynamic operation. lizing surplus power from renewable energy sources. Because renewable
- Proton exchange membrane electrolysis (PEMEC): Recently, the energy power cannot always be guaranteed to be used free of charge,
technical level has approached its completion stage, and its market and there are several operating cost items that can be affected by na­
share is rapidly increasing. Solid polymers conduct as electrolytes tional policy, it is needed to analyze their impact by setting up indi­
and membranes simultaneously and have a compact structure called vidual cost items without just assuming the operating cost as a value of
membrane electrode assembly (MEA), which is advantageous for the some proportional to the capital cost. The latter study deals with only
downsizing of electrolyzers. It typically operates at 50–80 ◦ C and simple results without the detailed process for economic analysis, and it
under a pressure of 200 bar. It can operate at a higher current density considered very limited cost factors. Thus, the results were insufficient
than AWE and create higher purity hydrogen gas. A key advantage of to assess the impact of various cost items on determining hydrogen
PEMEC is its flexible operation so that it shows a shorter response production cost for each system.
time than AWE [22]. However, using expensive catalyst materials The work’s prior objectives are to calculate the hydrogen production
such as platinum is the greatest challenge to solve. cost and compare the economic feasibility of mainly developed three
- Solid oxide electrolysis (SOEC): This technology can reach higher water electrolysis technologies, simultaneously. Since SOEC requires a
energy efficiency since less energy is needed to decompose water lot of energy to create a high-temperature condition, this paper presents
than AWE and PEMEC due to the high operating temperature. A two feasible systems by operating strategy. Especially, it is applied that
SOEC uses a ceramic membrane of yttria-stabilized zirconia (YSZ), results of the numerical simulation model of each electrolysis developed
which can operate in high-temperature conditions (600–1000 ◦ C) in previous studies. The hydrogen production rate for each electrolysis
[23]. It has been considered to be in the early-stage development in type was derived using system efficiency data of validated simulation
the laboratory. However, German manufacturer Sunfire is already results, not assuming commercial value. It is very meaningful in that it
selling small commercial products of reversible solid oxide cells. integrated economic study with energy analysis. Thus, it is one of the
Research also continues to improve its lifetime, which is one of the most prominent features compared to previous economic analysis
biggest problems of SOECs. Meanwhile, it requires a lot of energy to studies. By sensitivity analysis, the effect of various cost factors on the
create and maintain a high-temperature environment to perform hydrogen production cost of each electrolysis type was analyzed. Cost
high efficiency. Therefore, it is considered that the method of factors include investment, electricity cost, tax, capacity factor, and
coupling with the power plant that can supply waste heat will be others. This comprehensive analysis also complements the limitations of
competitive [24,25]. previous studies. Further, the probabilistic analysis considering un­
certainties was presented using Monte-Carlo analysis.
Currently, due to the electrolyzer system and renewable electricity
costs, green hydrogen is still supposed to be more expensive than con­ 2. Analysis method
ventional approaches. While the price of major materials for electro­
lyzers has been decreasing because of technological breakthroughs, the 2.1. H2 production cost estimation
whole system still has little profitability. Thus, comparative economic
studies of water electrolysis technologies need to continue Research can The method presented in the study by Nicita et al. [32] was used to
help establish strategies to reduce the price of green hydrogen produc­ estimate the hydrogen production cost using water electrolyzers. The
tion. Due to technical maturity, economic analysis studies of AWE and net present value (NPV) and the levelized cost of hydrogen (LCOH) were
PEMEC have been conducted mainly rather than SOEC. Genovese et al. evaluated based on the cash flow, which can be estimated by composing
[26] performed a comprehensive economic analysis for AWE based on expected revenue and expenditure items. In general, cash in the present
state-of-the-art technology of water electrolysis in the US and estimated is less valuable than the same amount of cash in the past. Consequently,
a base-case H2 production cost in 2009 of $5.20/kgH2 with a capacity of the NPV considers all the future net cash flows and assesses the project’s
50,000 kgH2/day. Ainscough et al. [27] estimated H2 production costs of final value over plant life at the present point in time [33]. The NPV was
$5.14/kgH2 for 1500 kgH2/day and $5.12/kgH2 for 50,000 kgH2/day in calculated by dividing the expected future cash flows from the plant
2013 in the US. Lee et al. [28] performed comparative analysis for on- operating by the discount rate and then summing them all, and it is
site hydrogen refueling stations using AWE, PEMEC, and SMR for H2 presented as follows:
production capacities of 30, 100, 300 Nm3/hr. The results show that
∑N ∑N
PEMEC produced hydrogen with the lowest cost of 16.54 $/kgH2 for 30 NPV =
CF n
n =
CF n
− I0 (2)
Nm3/hr capacity, while for 300 Nm3/hr capacity, SMR produced the n=0
(1 + r) n=1
(1 + r)n
lowest cost of 7.59 $/kgH2. Therefore, large-scale application SMR
would be economical, while PEMEC is acceptable for smaller-scale ap­ where N is the project period in years, CFn is the net cash flow at
plications. Lee et al. [29] conducted uncertainty analysis using Monte operating year n, I0 is the capital expenditure (CAPEX), and r is the
Carlo simulation with sensitivity analysis to estimate the hydrogen discount rate for calculating a change in cash value, which typically
production cost with a probabilistic assessment based on a previous applies 4–9% to energy industry facilities [34]. The discount rate reflects
study. the opportunity cost of the invested capital, the project’s riskiness, and
On the other hand, very few studies have been compared, including the inflation level [33].

3
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

0 or 1 that indicates whether oxygen is sold or not. Because hydrogen

and oxygen gas are generated at a mole ratio of 1:0.5 by water elec­
trolysis, it can be calculated that oxygen is coproduced 8 times the mass
of hydrogen.
The NPV is a measure of business feasibility. Given the required
earning rate (discount rate), a positive NPV means that revenue is
greater than expenditure and that the project is worth investing. Addi­
tionally, the payback period can be estimated when NPV becomes a
positive value, and NPV helps to choose more attractive projects from an
investor’s perspective. However, since NPV cannot fully consider the
difference in investment size between each project, financial metrics
called the internal rate of return (IRR) are analyzed. The IRR is the
discount rate that makes the final NPV zero and can be derived from the
equation as follows:
∑
N
CF n
=0 (6)
n=0
(1 + IRR)n

The calculated IRR value should be greater than the discount rate to
make revenue, and a project with a higher IRR can be expected to be
more competitive. Although NPV is the main factor for evaluating eco­
nomic feasibility, analyzing subsidiary factors such as IRR allows more
efficient comparison and decision making.
Based on the cost and revenue terms defined above, the LCOH was
calculated as follows:
TotalLifetimeCost
LCOH =
TotalLifetimeH2 Production
∑ ∑
CAPEX + Nn=1 OPEX n
(1+r)n
+ TR Nn=1 REVH n
(1+r)n
= ∑N MH2 (1− SRD)n (7)
n=1 (1+r)n

Fig. 2. (a) Model schematics of the AWE system and (b) numerical model where SRD is a degradation rate that accounts for a loss of hydrogen
validation results for AWE cell [36]. production efficiency. Generally, an electrolyzer stack is replaced after
the efficiency loss reaches 10%. Therefore, SRD for each electrolyzer
The cash flow of Eq. (2) can be replaced as follows: type was modified considering the operation lifetime and annual ca­
∑N pacity factor and applied to Eq. (5).
REV n − OPEX n
NPV = − CAPEX + (1 − TR) (3)
(1 + r)n
n=1
2.2. Sensitivity analysis
where REV n is the revenue from the product sale, OPEXn is the operating
Sensitivity analysis has been widely adopted for engineering simu­
expenditure, such as maintenance cost and electric cost, and TR is the
lation since it presents each input variable’s importance and helps to
tax rate imposed on net income. Dividing the cost and revenue parts
identify the critical parameters to the final outcomes. In this study,
allows Eq. (3) to be written as follows:
sensitivity analysis was conducted to determine the effect of variables on
∑N
OPEX n ∑N
REV n ∑N
REV n hydrogen production cost. It was investigated how the hydrogen pro­
NPV = − CAPEX − n − TR n+
(1 + r) (1 + r) (1 + r)n duction cost changes when a specific variable is increased or decreased
(4)
n=1 n=1 n=1
∑N by 20% is investigated. Exceptional items among the capital and oper­
OPEX n
+ TR ating costs and some economic indexes expected to impact were selected
(1 + r)n
n=1 as variables: electrolyzer stack, BOPs, labor, electricity, maintenance,
In particular, the last term means tax redemption, which excludes capacity factor, and tax rate.
some of the parts used as expenses to impose a tax on net income, not
sales revenue. This study assumed that there is no loan for the invest­ 2.3. Uncertainty analysis using Monte-Carlo simulation
ment cost while the manufacturer operates plants. Thus loan payments
and interest in debts were not considered. The Monte Carlo method (MCM) is defined as a method of evaluating
Once again, the REV n terms of Eq. (4) can be further specified as a measurement’s uncertainty. It relates to the propagation of the prob­
follows: ability distribution using a mathematical calculator model and applies
∑N ∑N REVHn + kREVOn ∑N MH2 (1 − SRD)n when several different input variables are used to derive the output
REVn
n = = PH2 result [35]. In particular, the economic evaluation combined with MCM
n=1 (1 + r) n=1 (1 + r)n n=1 (1 + r)n
allows reducing uncertainty and estimated future returns to be more
∑N 8MH2 (1 − SRD)n
+ k n=1 PO2 specified.
(1 + r)n In this study, MCM simulation was conducted using Microsoft Excel.
(5) The simulation model generated random numbers for each selected cost
item and derived the corresponding result value based on the designed
where REVHn is the revenue from the hydrogen sales, REVOn is the
mathematical modeling. According to the Gaussian probability distri­
revenue from the oxygen sales, MH2 is the annual production of
bution, the NORM.INV function was used to obtain arbitrary cost values
hydrogen in mass, SRD is a system degradation rate, PH2 and PO2 are the
considering the mean value and standard deviation. The adjusted stan­
selling price of hydrogen and oxygen, respectively, and k is a factor of
dard deviation is entered to limit the variation in the cost value to a

4
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

Fig. 3. (a) Model schematics of the PEMEC system and (b) numerical model
validation results for PEMEC cell [38].

particular range. After many repetitions of the calculation above, sta­

tistical analysis is conducted from the aggregated output data series. In
this study, the calculation was repeated 2000 times. The RAND function
was used to specify random probability values following the normal
distribution for each calculation.

3. System modeling and case analysis

Four system cases were considered to evaluate the economic char­

acteristics of a grid-connected hydrogen production plant: the AWE
system, PEMEC system, SOEC system with electric heaters (E.H), SOEC
system with waste heat sources (W.H). Following sections describe nu­ Fig. 4. Model schematics of the SOEC system (a) with electric resistance
merical models of each water electrolysis system. Based on the simula­ heaters, (b) with waste heat source, and (c) numerical model validation results
for reversible SOFC cell [39].
tion results of each system model developed in the previous study
[36–39], reference operation points were selected and system effi­
ciencies were derived. Obtained operating efficiencies were used to
Table 1
calculate the hydrogen production rate of each system. Analysis of the water electrolysis system for the derivation of the operating cost
index.

3.1. AWE system Case AWE PEMEC SOEC (E. SOEC (W.
H) H)

Fig. 2 (a) shows model schematic of the AWE system. A numerical Current density [A/cm2] 0.40 1.20 0.67 0.67
Cell voltage [V/cell] 1.88 1.98 1.26 1.26
model of an AWE system was developed using Aspen Plus® [36]. The
Supplied power [MWel] 1.0 1.0 1.0 1.0
AWE system consists of electrolysis stack and BOP devices including System efficiency (LHV base) 65.75 62.21 61.4 87.6
heaters, pumps, phase separators, and EHP cooling system. While [%]
operating the AWE stack, an excess flow of aqueous electrolyte solution H2 production [Nm3/hr] 109.95 104.01 102.66 146.47
is circulated to both anode side and cathode side. Therefore, BOP de­
vices for the electrolyte circulation and temperature control are required model can be referred to in our previous work [36,37]. The net voltage
for the anode side and the cathode side stream, respectively. of the AWE stack was calculated as the sum of the reversible voltage and
The stack incorporated the following zero-dimensional model to activation overvoltage and ohmic overvoltage as follows:
simulate electrochemical reactions. The specific description of this

5
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

Table 2 Table 4
Cost factors and values comprising the operating cost. Categorization of cost items and hydrogen production cost of each electrolyzer
Items Value Unit
type.
Items AWE PEMEC SOEC (E. SOEC
Overall Discount rate (r) 8 %
H) (W.H)
Economic Income tax rate (TR) 30 %
Index Capital Cost Electrolyzer 340,000 420,000 520,000 520,000
Plant Operating Plant lifetime 20 Years [$] Stack
Index Stack lifetime AWE [45] 80,000 Hours BOP
PEMEC 40,000 Hours Power supply 198,225 198,225 198,225 198,225
[45] Water 87,082 63,541 87,082 87,082
SOEC 20,000 Hours circulation
[46] Hydrogen 83,880 83,880 83,880 83,880
Operating hour 12 Hours/day processing
System degradation AWE 1.3 %/10,000 Cooling 28,678 21,517 28,678 39,203
rate (SRD) hr Electric heater – – 15,000 –
PEMEC 2.7 %/10,000 Other BOPs 6000 6000 6000 6000
hr Total 743,865 793,163 938,865 934,390
SOEC 5.9 %/10,000 Operating Labor 36,575 36,575 36,575 36,575
hr Cost Electricity 292,000 292,000 292,000 292,000
Plant Operating Electric cost [34] 66.7 $/MWh [$/year] Water 8,649 8,182 8,076 11,522
Cost Labor [28] 36,575 $, full time KOH 480 – – –
Deionized water [47] 0.01 $/kg Service and 14,877 15,863 18,777 18,688
KOH [47] 2.92 $/kg Maintenance
Service and Maintenance [28] 2 % of CAPEX Other operating 7439 7932 9389 9344
Other operating cost [28] 1 % of CAPEX cost
Total 360,020 360,552 364,817 368,129
Hydrogen Production rate [kgH2/ 86,488 81,819 80,760 115,221
year]
Table 3 LCOH [$/kgH2] 7.599 8.550 10.156 7.155
Cost values comprising the capital cost of water electrolysis plants.
Items Value Unit
( ) ( )
Stack Cost AWE [45] 340,000 $/1 MW RT j RT j
PEMEC [45] 420,000 system Vact = sinh− 1
+ sinh− 1
(10)
F αa j0,a F αc j0,c
SOEC [46] 520,000
BOP Cost Power supply 27,556
[44] Deionized water Separator tank 10,000 where j0 and α represent the exchange current density and the charge
circulation Circulation 5,500 transfer coefficient, respectively. Subscripts a and c mean anode side and
pump
cathode side, respectively. Finally, the ohmic overvoltage was calcu­
Piping 7,600
Valves, control 7,700 lated from the sum of each internal resistance element as follows:
Hydrogen Dryer 13,900
Vohm = I × (Rcathode + Ranode + RKOH + Rmem ) (11)
Processing Hydrogen 5,300
separator
Tubing 3,800 where subscripts cathode, anode, KOH, and mem mean cathode side
Valves, control 5,200 electrode, anode side electrode, KOH aqueous electrolyte, and mem­
Cooling Processing Heat exchanger 9,000 brane, respectively. Resistance factors were calculated based on the
Cooling pump 15,000
electrical or ion conductivity of each component, and especially KOH
Dry cooler 4,000
Valves, control 3,000 resistance took into account the influence of generated bubbles due to
Electric resistance heater [43] 7,500 the electrolyzer operation. The developed AWE numerical model has
Other BOPs 6,000 been validated by comparing the I-V polarization curve with the
experiment data which was conducted by KIER as shown in Fig. 2 (b).
The AWE system efficiency was calculated by dividing the heating
Vcell = Erev + Vact− c + Vact− a + Vohm (8)
value of produced hydrogen gas with the system required power as
follows:
Erev is the reversible voltage, Vact− c ,Vact− a is the activation voltage for the
cathode and anode, and Vohm is the ohmic overvoltage. The reversible nH2 ,prod ∙LHVH 2
ηsystem = (12)
voltage was calculated by the Nernst equation. Wsystem
( )
RT (P − PH2O )1.5 P*H2O where LHVH2 is the lower heating value of the produced hydrogen and
(9)
◦
Erev = Erev + ln
2F PH2O Wsystem is the total power supply for system operation, including the
stack and BOPs.
The first term Erev is the standard potential and represents the effect
of operating temperature. The second term represents the effect of
pressure, which was calculated as a function of the operating pressure 3.2. PEMEC system
and the water vapor partial pressure. P, T, R, and Fare the operating
pressure, temperature, universal gas constant, and Faraday constant, A numerical model of a PEMEC system is also developed using Aspen
respectively. The activation overvoltage is required while an electro­ Plus®. Fig. 3 (a) shows the system schematic of the PEMEC system. The
chemical reaction breaks out of the equilibrium state. That is, while the PEMEC is classified as low-temperature electrolysis technology along
reversible voltage sets an equilibrium state to get ready for initiating the with AWE. Thus it has a similar BOP configuration to AWE. However,
electrochemical reaction, the activation overvoltage is requested to the most obvious difference seen in the schematic is that there is no
trigger the reaction. The Butler-Volmer equation is generally used to system for circulating and re-supplying pure water on the cathode side
calculate the activation overvoltage as follows: of the PEMEC. In the AWE, the aqueous KOH solution electrolyte cir­
culates, and it needs to maintain the electrolyte concentration balance

6
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

Fig. 5. Ratio of cost items that contribute to the calculated levelized cost of hydrogen configuration for each electrolysis system type. The numbers above each bar
indicate the levelized cost of hydrogen value.

on both sides. Unlike AWE, the PEMEC which circulates pure water may
Vcell = VN + j × ASR (15)
supply water only on the anode side where the decomposition reaction
occurs.
where, VN , j, and ASR is Nernst voltage, average current density, and
The net voltage of the PEMEC stack was calculated as follows:
area specific resistance, respectively. Nernst voltage was calculated as
Vcell = Erev + Vact− c + Vact− a + Vohm + Vcon (13) follows:
√̅̅̅̅̅̅̅̅
The electrochemical reaction model is also similar to the AWE. ΔG RT PH2 PO2
VN = + ln (16)
However, the PEMEC further considers a concentration overvoltage 2F 2F PH2O
term, Vcon , that is increased the influence in the high current density
range. It is caused by the movement of oxygen gas bubbles in the anode where, ΔG is absolute Gibbs energy change. Fig. 4 (c) shows model
component layers. The concentration overvoltage can be expressed by simulation results compared to the experimental data from the study of
Nernst equation as follows: Jensen et al. [41].
[ ] The system efficiency calculation was also similar to the previous
RT aO2 aHcc2 aHac2 O two water electrolysis technologies. Therefore, efficiency was calculated
Vcon = ln ac + 2ln − 2ln (14)
4F aOaf2 aHcf2 aHaf2 O as follows:
nH2 ,prod ∙LHVH 2
where superscript a represents the activity of each component and ηsystem = (17)
Wsystem + Wboil
subscripts means model boundary such as catalyst layer and flow
channel on both anode and cathode sides. The specific description of the nH2 ,prod ∙LHVH 2
model can be referred to in our previous work [38]. Developed PEMEC ηsystem = (18)
Wsystem
numerical model has been validated by comparing the I-V polarization
curve with the experiment data which was conducted by KIER as shown where, Eq. (17), which contains the heat required to generate high-
in Fig. 3 (b). temperature steam, corresponds to the SOEC (E.H) case, and Eq. (18)
The PEMEC system efficiency was also calculated by dividing the corresponds to the SOEC (W.H) case.
heating value of produced hydrogen gas with the system required
power, as shown in Eq. (12).
3.4. Cases summary

3.3. SOEC system All system cases were assumed to be 1 MWel class plants. This is
equivalent to the state-of-the-art stack unit size of low-temperature
A numerical model of a SOEC system is referred to the operation data electrolysis technologies [42] and it can have the advantage of modu­
of SOEC mode among the previous work [39] that developed reversible larizing to the size of a large renewable energy plant. Table 1 shows the
SOFC system model. This model was developed based on lumped models calculated operation point and system efficiency. Based on the system
in EBSILON® Professional software. Fig. 4 (a) and (b) show two different supply electric power, 1 MWel, and each system efficiency, the reference
system schematics of the SOEC system. The SOEC that is classified as annual hydrogen production rates were derived.
high-temperature electrolysis technology requires heating up to about
750 ◦ C. In this regard, the former schematic uses electric resistance 3.5. Cost data
heaters and the latter one uses heat exchangers with an external waste
heat source. Waste heat is assumed to be used without cost under the The annual operating cost indexes are summarized in Table 2. These
situation the SOEC system is coupled with a plant generating high- indexes were applied to determine the OPEX of Eq. (3). The plant
temperature waste heat as a by-product. Nuclear power plants can be operating lifetime was set to 20 years, and stacks were completely
an appropriate example [40]. replaced for each specified stack lifetime. According to the IRENA report
The net voltage of the SOEC stack was calculated as follows: [42], it is said that about 4000 h of operation per year is required for

7
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

Table 5
Internal ratio of return (IRR) and payback period (PBP) analysis for each elec­
trolysis system type according to the hydrogen sales price.
Hydrogen 7.50 $/kgH2 8.00 $/kgH2 9.00 $/kgH2
sales price
IRR PBP IRR PBP IRR PBP
[%] [year] [%] [year] [%] [year]

AWE 6.613 – 12.816 9.61 22.597 5.27

(10th) (6th)
PEMEC – – 7.361 – 13.036 10.91
(11th)
SOEC (E.H) – – – – – –
SOEC (W.H) 12.699 11.01 18.636 6.66 29.209 3.26
(12th) (7th) (4th)

electrolyte solutions. Maintenance and other operating costs were

assumed to be 2% and 1% of the capital cost, respectively [28].
The initial investment costs of stacks and all BOPs were summarized
in Table 3. Specific BOP cost values for the 1 MWel system presented in
the NREL report [44] were applied, constituting the CAPEX of Eq. (3) for
each case. Moreover, additional hydrogen processing part costs were
applied to the oxygen processing part when selling the coproduced ox­
ygen gas.
The main assumptions established for economic analysis in this study
are summarized as follows:

- 1 Mwel hydrogen production plants are compared for each electrol­

ysis type.
- The plants were operated for 12 h every day (50% capacity factor,
4,380 h per year).
- The plant lifetime is 20 years, the first year is the construction period,
and there is no profit.
- Each stack is completely replaced at the end of its lifetime, and the
degradation rate of each electrolyzer type is applied, assuming 10%
efficiency loss occurs during the lifetime.
- One full-time worker manages the plants, and the electric charge is
calculated as industrial electricity (66.7 $/MWh) in Korea.
- The discount rate (r) and tax rate (TR) are applied at 8% and 30% for
the net present value calculation, respectively.
- The price is calculated as 0.054 $/kgO2 when selling coproduced
oxygen gas [47].

4. Analysis results and discussion

4.1. Analysis results for the reference condition

Technoeconomic analysis of each electrolyzer type was conducted

for the 1 MWel hydrogen production plant. Table 4 shows categorized
capital costs and operating costs, annual hydrogen production rates, and
hydrogen production costs for AWE, PEMEC, SOEC (E.H), and SOEC (W.
Fig. 6. Net present value (NPV) variation for each electrolysis system type H). It should be noted that the operation cost and hydrogen production
according to the hydrogen sales price of (a) 7.5 $/kgH2, (b) 8.0 $/kgH2, and (c) rate are reference values that do not apply to the discount rate or
9.0 $/kgH2. degradation rate, while the LCOH is the value that includes all variables
throughout the entire plant lifetime.
green hydrogen production by water electrolysis to be competitive Capital costs were obtained by combining the price of BOP devices
against fossil-based hydrogen. Therefore, the capacity factor was set to based on the NREL report [44] and the stack price according to each
be 50% as a reference condition, which means that the plant operates 12 system configuration presented in the previous section. For operating
h a day and 4380 h a year. The system degradation rate per 10,000 h was costs, labor and electric costs were identical under the assumption that
calculated to decrease by 10% over the stack lifetime considering the equivalent 1 MWel plants were compared. Water is a material that is
variable capacity factors. The electric cost was calculated based on the decomposed to produce hydrogen gas, and KOH is applied only to AWE
industrial electric cost in Korea, which is approximately 66.7 $/MWh. systems.
The labor cost was calculated based on the assumption that one full-time The hydrogen production rate was calculated as the amount that
worker is employed. Water was assumed to use 10 kg to produce 1 kg of each plant could produce considering the BOP power consumption and
hydrogen, considering some losses included during the hydrogen puri­ operation efficiency when 1 MWel power was supplied, and it was higher
fication process, and the total cost was calculated with the price pre­ in the SOEC (W.H) > PEMEC > AWE > SOEC (E.H) order. The hydrogen
sented in Table 2. KOH was used only in the AWE system to make production rate of SOEC (W.H) showed the largest value of 115,221

8
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

Fig. 7. Sensitivity analysis results for each system case of (a) AWE, (b) PEMEC, (c) SOEC (E.H), (d) SOEC (W.H).

Fig. 8. Levelized cost of hydrogen variation for each electrolysis system type according to (a) electric cost and (b) capacity factor changes.

kgH2/year because it has the advantages of high system efficiency while The water portion of the AWE case included the KOH cost, accounting
saving energy for generating water vapor. On the other hand, SOEC (E. for approximately 2% of the LCOH. The capital costs that cover the stack
H) had a reduced net power supply that contribute to hydrogen pro­ and BOP costs contributed approximately 17%. Although there are dif­
duction due to a large amount of electric heater power for water vapor. ferences in each electrolysis type, the difference is not noticeable due to
As a result, it showed the lowest value of 80,760 kgH2/year. AWE and other cost items with a relatively significant influence. The most obvious
PEMEC case produced 86,488 kgH2/year and 80,760 kgH2/year, difference is seen in the electricity and stack replacement costs. In
respectively. The tendency of the LCOH also followed the order of particular, the electric cost accounts for at least half of the LCOH for­
hydrogen production rate. The LCOH was obtained by comparing the mation. It follows the stack operation efficiency trend and shows the
cost spent during the plant lifetime with the amount of produced lowest proportion in the SOEC cases, especially when using waste heat.
hydrogen during the lifetime, as shown in Eq. (7). The SOEC (W.H) has The stack replacement costs are naturally affected by the stack price.
the lowest price of 7.155 $/kgH2, followed by AWE, PEMEC, SOEC (E.H) Furthermore, since the SOEC stack is expensive and has a short lifetime,
with 7.599 $/kgH2, 8.550 $/kgH2, and 10.156 $/kgH2, respectively. it shows a remarkable influence compared to the AWE and PEMEC cases.
SOEC (W.H) and AWE systems can be chosen as inexpensive It is possible to draw a cumulative cash flow diagram with NPV data
hydrogen production methods due to the lowest investment cost in the calculated by Eq. (2) and perform a profitability analysis of the project.
AWE case and the lowest electrical power requirement in the SOEC (W. Fig. 6 shows the NPV variation for each electrolysis system type by three
H) case. However, the assumption that the SOEC (W.H) system can hydrogen sales price cases. Commonly, each diagram shows the change
utilize waste heat sources without a charge to operate in high- over the project period of 20 years after the initial investment (CAPEX)
temperature conditions must be satisfied. Otherwise, if some power expenditure incurred in the first year, assuming the construction period.
has to be used to meet the operating conditions, the LCOH increases The parts where the graph drops indicate the replacement cost expen­
since it cannot produce sufficient hydrogen to satisfy high investment diture due to the stack operating lifetime. Therefore, the SOEC cases,
costs such as the SOEC (E.H) case. which have the shortest stack lifetime, show the most replacement
The total cost for the overall project period was itemized, and the points of 4 times during 20 years under the reference condition of a 50%
ratio of the LCOH composition was computed, which is shown in Fig. 5. capacity factor. The point at which NPV is completely converted to a

9
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

Fig. 9. Comparison of each levelized cost of hydrogen value in a scenario assuming government funding and coproduced oxygen sales case.

positive value as revenue continues to arise is referred to as the payback influence in the SOEC system.
period (PBP). Thus, it is possible to roughly determine the project’s Fig. 8 shows the LCOH for a wider range of electric costs and capacity
profitability by whether the curve crosses the zero line of the y-axis. That factors. Generally, the LCOH increased as the electric cost rose, as seen
is, if NPV is negative at the end of the project, profitability cannot be from the sensitivity analysis. One interesting thing from Fig. 8 (a) is that
expected. Meanwhile, calculating the internal ratio of return (IRR) by while the SOEC system cases were able to compensate high investment
Eq. (6) gives a more specific profitability comparison. It must be costs by operating with high electric efficiency, the benefits were
calculated to be higher than the set discount rate of 8% to make a net weakened by lower electric costs. Eventually, SOEC (W.H) case became
income. It is more favorable to have a project with a higher IRR. The IRR more expensive than the AWE case under an electric cost of 25 $/MWh.
and PBP were calculated to compare the competitiveness of each case for In contrast, higher electric costs made a more considerable difference
the assumption of three different hydrogen prices. The specific IRR and from other cases. Fig. 8 (b) shows that the LCOH increased exponentially
PBP values of each case are shown in Fig. 6 and summarized in Table 5. when the capacity factor decreased since it became more difficult to
The SOEC (W.H) case was profitable for all three sales price cases and recover the investment cost during a plant lifetime. According to the IEA
shows net income since the 9th project year even at the lowest selling report in 2018 [48], the average capacity factors of onshore and offshore
price, whereas the SOEC (E.H) case never reach PBP. The AWE case and wind plants are 23–44% and 29–52%, respectively, and solar PV is only
the PEMEC case started to be profitable from 8.0 $/kgH2 and 9.0 $/kgH2, 10–21%. Consequently, when designing water electrolysis coupled with
respectively, and reached the PBP at the 6th and 11th project years when renewable energy sources, which are essential for a green energy soci­
selling at the highest price case of 9.0 $/kgH2. When hydrogen was sold ety, the impact of a low-capacity factor on the LCOH must be considered.
at the lowest price, the IRR of three cases except for the SOEC (W.H) case Fig. 9 shows calculated LCOH values in a scenario assuming gov­
was lower than 8%, the reference discount rate, or not calculated. This ernment funding and coproduced oxygen sales cases. The baselines for
also confirmed that three cases do not make a profit. In the 8.0 $/kgH2 each electrolysis indicated by the dotted line show the reference LCOH
sales case, the IRR of AWE was 8.853 and slightly exceeded the reference shown in Table 4. If 8 kg of coproduced oxygen and 1 kg of generated
discount rate. The IRR of AWE, PEMEC, SOEC (W.H) was calculated as hydrogen was also sold, the LCOH values were reduced by approxi­
14.341, 11.826, 27.380, respectively, in the 8.5 $/kgH2 sales case. The mately 0.25 $/kgH2, although there were a few differences between each
SOEC (E.H) had an IRR of 6.106, which cannot make net revenue during electrolysis. As confirmed by the sensitivity analysis, there was a more
the project periods, as also shown by the PBP analysis. substantial impact on reducing LCOH in the order of electric cost, tax
rate, and labor cost. Case 3, which does not electricity cost, showed a
lower LCOH than Case 8, where electricity cost and taxes are supported
4.2. Sensitivity analysis
by half. The lowest LCOH for Case 3 and Case 8 was AWE with 2.60
$/kgH2 and SOEC (W.H) with 4.34 $/kgH2, respectively. If electricity
In this study, a sensitivity analysis was conducted to determine the
costs and tax imposition can be completely ignored, AWE, PEMEC, SOEC
effect on hydrogen production cost by varying the value of specific cost
(E.H), and SOEC (W.H) can produce hydrogen at 1.82, 2.28, 3.34, and
items by ±20%. The sensitivity analysis results for each electrolysis
2.37 $/kgH2, respectively.
system case are shown in Fig. 7. Although there are differences in
magnitude, the overall trend is similar for all cases. The electric cost had
the most dominant effect, followed by the tax rate. However, the third 4.3. Uncertainty analysis
factor was the capacity factor for the AWE and PEMEC systems while the
stack cost for the SOEC system. Due to the relatively expensive cost and Contrary to the SA in the previous section that analyzed changes in
more frequent replacement with a short lifetime, the stack cost had more LCOH by only varying individual cost items, the uncertainty analysis

10
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

Fig. 10. Probability histograms for levelized cost of hydrogen of (a) AWE, (b) PEMEC, (c) SOEC (E.H), (d) SOEC (W.H), and (e) cumulative probability curve for
comparison of electrolysis system cases.

Fig. 11. Probability histograms for the net present value of (a) AWE, (b) PEMEC, (c) SOEC (E.H), (d) SOEC (W.H) when produced hydrogen is sold for 9 $/kgH2. The
dark part of each histogram indicates the probability that the project does not reach the payback period.

11
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

considers the comprehensive changes in the selected cost items and capacity factor in the AWE and PEMEC cases showed different results.
analyzes the stochastic results. The selected cost items were the same as Due to the relatively expensive cost and more frequent replacement with
those in SA as stack, BOPs, labor, electricity, maintenance, tax, and a short lifetime, the stack cost had more influence in the SOEC system.
capacity factor. The program generated random values within ±20% of The LCOH and profitability of each electrolysis technology can be
each reference value and collected LCOH and NPV data with 2000 analyzed probabilistically through the MCM. While calculating 2000
calculations. times using randomly adjusted cost items, each LCOH was most likely
Fig. 10 shows the probability for the LCOH of each electrolysis sys­ calculated as 7.60 $/kgH2 with AWE, 8.35 $/kgH2 with PEMEC, 10.10
tem case. Each bar is divided into $0.05, and its value is the proportion $/kgH2 with SOEC (E.H), and 7.00 $/kgH2 with SOEC (W.H). When
of the data in each range to the total LCOH data. Therefore, the largest selling generated hydrogen at 9 $/kgH2, the SOEC (E.H) cannot be ex­
value indicates the range of costs formed with the highest probability pected to be profitable, while AWE, PMEEC, and SOEC (W.H) were ex­
considering several variable cost items. The simulation results showed pected to have positive probabilities of 99.7%, 78.8%, and 99.9%,
that AWE and PEM cases mainly produce hydrogen in the price range of respectively.
7.60–7.65 $/kgH2 and 8.35–8.40 $/kgH2, while SOEC (E.H) and SOEC
(W.H) cases produce hydrogen with the highest probability at price CRediT authorship contribution statement
ranges of 10.10–10.15 $/kgH2 and 7.00–7.05 $/kgH2, respectively.
Compared to the LCOHs for the previous data with fixed cost items, Dohyung Jang: Methodology, Software, Validation, Resources,
stochastic data of uncertainty analysis were predicted to be slightly Writing – original draft. Jaedong Kim: Validation, Resources, Funding
higher. By expanding the range to ±20% based on the central value, it acquisition. Dongmin Kim: Validation, Resources, Funding acquisition.
was predicted that the AWE and PEMEC cases generate in the range of Won-Bi Han: Validation, Resources, Funding acquisition. Sanggyu
8.00–8.20 $/kgH2 and 8.59–8.86 $/kgH2, and SOEC (E.H) and SOEC (W. Kang: Conceptualization, Formal analysis, Writing – review & editing,
H) cases are in the range of 7.81–8.17 $/kgH2 and 9.99–10.34 $/kgH2, Supervision, Funding acquisition.
respectively. These trends for the LCOH of each electrolysis system case
can be compared through the cumulative curve shown in Fig. 10(e).
Declaration of Competing Interest
Fig. 11 shows the probability for the NPV of each electrolysis system
case when hydrogen is sold for 9 $/kgH2. For three system cases, except
The authors declare that they have no known competing financial
for SOEC (E.H) case, it was predicted that there is a good chance of
interests or personal relationships that could have appeared to influence
having a positive NPV value. That is, these cases have profitability. The
the work reported in this paper.
dark side of each histogram indicates the probability that the project
does not reach the PBP. The probability that the SOEC (E.H) case is
Acknowledgments
profitable was zero, and the PEMEC case had a 21.2% possibility of
negative NPV, which cannot be ignored. Meanwhile, compared to the
This work was supported by the Korea Gas Corporation (KOGAS)
AWE case, the SOEC (W.H) case was expected to have larger NPV values,
Research Institute and by the Hydrogen Energy Innovation Technology
while both systems have a very small probability of having a negative
Development Program of the National Research Foundation of Korea
NPV, less than 0.2%. The expected range was also widely formed due to
(NRF), funded by the Korean government (Ministry of Science and ICT
relatively large cost items affected, such as investment cost and
[MSIT]) (NRF2019M3E6A1064287).
hydrogen production rate.

References
5. Conclusions
[1] Redl C, Hein F, Buck M, Graichen P, Jones D. The European Power Sector in 2019:
In this study, the LCOH and the NPV of each water electrolysis sys­ Up-to-Date Analysis on the Electricity Transition. Tech. rep. Agora Energiewende
tem case were calculated to analyze and compare economic feasibility. and Sandbag; 2020.
[2] Breeze P. Power system energy storage technologies. Academic Press; 2020.
Sensitivity analysis and Monte Carlo analysis were conducted to inves­ [3] López González E, Isorna Llerena F, Silva Pérez M, Rosa Iglesias F, Guerra Macho J.
tigate the effect of varying cost items and uncertainty. Energy evaluation of a solar hydrogen storage facility: comparison with other
The cost required to produce 1 kg of hydrogen in the operation of the electrical energy storage technologies. Int J Hydrogen Energy 2015;40(15):
5518–25.
1 MWel plant of each electrolysis type was calculated as 7.60 $ with
[4] Dawood F, Anda M, Shafiullah GM. Hydrogen production for energy: an overview.
AWE, 8.55 $ with PEMEC, 10.16 $ with SOEC (E.H), and 7.16 $ with Int J Hydrogen Energy 2020;45(7):3847–69.
SOEC (W.H). SOEC combined with waste heat case was found to be the [5] Niaz S, Manzoor T, Pandith AH. Hydrogen storage: material, methods and
most economical. perspectives. Renew Sustain Energy Rev 2015;50:457–69.
[6] Hassanpouryouzband A, Joonaki E, Vasheghani Farahani M, Takeya S, Ruppel C,
According to the NPV analysis, comparing the cases in which Yang J, et al. Gas hydrates in sustainable chemistry. Chem Soc Rev 2020;49(15):
hydrogen is sold for 7.5 $/kgH2, 8.0 $/kgH2, 9.0$/kgH2, the SOEC (W.H) 5225–309.
case made a profit on all cases, while the AWE and PEMEC cases started [7] Hassanpouryouzband A, Joonaki E, Edlmann K, Haszeldine RS. Offshore geological
storage of hydrogen: is this our best option to achieve net-zero? ACS Energy Lett
to generate revenue from 8.0 $/kgH2 and 9.0 $/kgH2, respectively. 2021;6(6):2181–6.
However, the SOEC (E.H) case did not convert NPV to positive value [8] Jeon J, Kim SJ. Recent progress in hydrogen flammability prediction for the safe
even though it sold at the highest price. energy systems. Energies 2020;13:6263.
[9] Najjar YSH. Hydrogen safety: the road toward green technology. Int J Hydrogen
DOE has set a target supply price of 10 $/kgH2 for hydrogen from Energy 2013;38(25):10716–28.
water electrolysis, and the corresponding target production cost is 5 [10] Hübert T, Boon-Brett L, Palmisano V, Bader MA. Bader MA Developments in gas
$/kgH2 [49]. According to the hydrogen production price by the support sensor technology for hydrogen safety. Int J Hydrogen Energy 2014;39(35):
20474–83.
scenario calculated in this study, the electric costs must be reduced by [11] Derwent RG. Hydrogen for heating: atmospheric impacts. Department for Business,
less than half to reach the target price, and if both electric cost and taxes Energy and Industrial Strategy Research Paper 2018:21.
are reduced by half, the price is less than 5 $/kgH2 except for the SOEC [12] Popa ME, Segers AJ, van der Gon HD, Krol MC, Visschedijk AJH, Schaap M, et al.
Impact of a future H2 transportation on atmospheric pollution in Europe. Atmos
(E.H) case. From a long-term perspective, it must be accompanied by a
Environ 2015;113:208–22.
significant drop in investment costs. [13] Wild O, Fiore AM, Shindell DT, Doherty RM, Collins WJ, Dentener FJ, et al.
According to the sensitivity analysis, electric cost and tax rate were Modelling future changes in surface ozone: a parameterized approach. Atmos
the most dominant and secondary effective variables, respectively, in Chem Phys 2012;12(4):2037–54.
[14] van Ruijven B, Lamarque J-F, van Vuuren DP, Kram T, Eerens H. Emission
determining the LCOH commonly for every case. However, the tertiary scenarios for a global hydrogen economy and the consequences for global air
effective variable, which is the stack cost in the SOEC case, and the pollution. Global Environ Change 2011;21(3):983–94.

12
D. Jang et al. Energy Conversion and Management 258 (2022) 115499

[15] Vogel B, Feck T, Grooß JU, Riese M. Impact of a possible future global hydrogen [33] Žižlavský O. Net present value approach: method for economic assessment of
economy on Arctic stratospheric ozone loss. Energy Environ Sci 2012;5:6445–52. innovation projects. Procedia-Soc Behav Sci 2014;156:506–12.
[16] Wang D, Jia W, Olsen SC, Wuebbles DJ, Dubey MK, Rockett AA. Impact of a future [34] Jeong S, Choi NH, Moon C, Lim H. Economic feasibility analysis for P2G using PEM
H2-based road transportation sector on the composition and chemistry of the water electrolysis. Trans Korean Hydrogen New Energy Soc 2017;28:231–7.
atmosphere–Part 2: Stratospheric ozone. Atmos Chem Phys 2013;13:6139–50. [35] Broday EE, Ruivo CR, Silva da MG. The use of Monte Carlo method to assess the
[17] Jacobson MZ. Effects of wind-powered hydrogen fuel cell vehicles on stratospheric uncertainty of thermal comfort indices PMV and PPD: benefits of using a measuring
ozone and global climate. Geophys Res Lett 2008;35(19). set with an operative temperature probe. J Build Eng 2021;35.
[18] Abe JO, Popoola API, Ajenifuja E, Popoola OM. Hydrogen energy, economy and [36] Jang D, Cho HS, Kang S. Numerical modeling and analysis of the effect of pressure
storage: review and recommendation. Int J Hydrogen Energy 2019;44(29): on the performance of an alkaline water electrolysis system. Appl Energy 2021;
15072–86. 287.
[19] Keipi T, Tolvanen H, Konttinen J. Economic analysis of hydrogen production by [37] Jang D, Choi W, Cho HS, Cho WC, Kim CH, Kang S. Numerical modeling and
methane thermal decomposition: comparison to competing technologies. Energy analysis of the temperature effect on the performance of an alkaline water
Convers Manag 2015;159:264–73. electrolysis system. J Power Sources 2021;506.
[20] Topler J, Lehmann J. Hydrogen and fuel cell: technologies and market [38] Ahn JW, Jang DH, Kang S. Analysis of the effect of pressure on the performance of
perspectives. Springer; 2015. polymer electrolyte membrane water electrolysis system. Conference of the Korean
[21] Godula-Jopek A. Hydrogen production by electrolysks. KGaA: Wiley-VCH Verlag Hydrogen & New Energy Society (KHNES), Daejeon, South Korea, 27-28 May,
GmbH & Co; 2015. 2021, pp. 233.
[22] Stolten D, Emonts B. Hydrogen science and engineering: materials, processes, [39] Giap V-T, Kang S, Ahn KY. High-efficient reversible solid oxide fuel cell coupled
systems and technology. KGaA: Wiley-VCH Verlag GmbH & Co; 2016. with waste steam for distributed electrical energy storage system. Renew Energy
[23] Wendel CH, Braun RJ. Design and techno-economic analysis of high efficiency 2019;144:129–38.
reversible solid oxide cell systems for distributed energy storage. Appl Energy [40] Kim JS, Boardman RD, Bragg-Sitton SM, Jacobson MZ. Dynamic performance
2016;172:118–31. analysis of a high-temperature steam electrolysis plant integrated within nuclear-
[24] Toklu E, Coskun Avci A, Kaygusuz K, Gur M. A research on hydrogen production renewable hybrid energy systems. Appl Energy 2018;228:2090–110.
from industrial waste heat by thermal water splitting. Int J Hydrogen Energy 2016; [41] Jensen SH, Larsen PH, Mogensen M. Hydrogen and synthetic fuel production from
41(24):10071–9. renewable energy sources. Int J Hydrogen Energy 2007;32(15):3253–7.
[25] Wang F, Wang L, Ou Y, Lei X, Yuan J, Liu X, et al. Thermodynamic analysis of solid [42] IRENA. Green hydrogen cost reduction. Abu Dhabi, AE: International Renewable
oxide electrolyzer integration with engine waste heat recovery for hydrogen Energy Agency; 2020.
production. Case Stud. Thermal Eng 2021;27:101240. [43] Harvego EA, O’Brien JE, McKellar MG. System evaluation and life-cycle cost
[26] Genovese J, Harg K, Paster M, Turner J. Current state-of-art hydrogen production analyses for high-temperature electrolysis hydrogen production facilities. USA:
cost estimate using water electrolysis. NREL Tech. Rep.; 2009. Idaho National Laboratory; 2012.
[27] Ainscough C, Peterson D, Miller E. Hydrogen production cost from PEM [44] Ruth MF, Mayyas AT, Mann MK. Manufacturing competitiveness analysis for PEM
electrolysis. USA: Department of Energy; 2014. and alkaline water electrolysis systems. In Proceedings of the 2017 Fuel Cell
[28] Lee B, Chae H, Choi NH, Moon C, Moon S, Lim H. Economic evaluation with Seminar and Energy Expo, Long Beach Conferation Center, Long Beach, CA, USA,
sensitivity and profitability analysis for hydrogen production from water 2017.
electrolysis in Korea. Int J Hydrogen Energy 2017;42:6462–71. [45] Nguyen T, Abdin Z, Holm T, Merida W. Grid-connected hydrogen production via
[29] Lee B, Heo J, Choi NH, Moon C, Moon S, Lim H. Economic evaluation with large-scale water electrolysis. Energy Convers Manag 2019;200:112108.
uncertainty analysis using a Monte-Carlo simulation method for hydrogen [46] Zauner A, Bohm H, Rosenfeld DC, Tichler R. Innovative large-sacle energy stroage
production from high pressure PEM water electrolysis in Korea. Int J Hydrogen technologies and Power-to-Gas concepts after optimization. D7.7 Analysis on
Energy 2017;42(39):24612–9. future technology options and on techno-economic optimization 2019:1–89.
[30] Ferrero D, Gamba M, Lanzini A, Santarelli M. Power-to-gas hydrogen: techno- [47] Meunier N, Chauvy R, Mouhoubi S, Thomas D, Weireld GD. Alternative production
economic assessment of processes towards a multi-purpose energy carrier. Energy of methane from industrical CO2. Renewable Energy 2020;146:1192–203.
Procedia 2016;101:50–7. [48] Average annual capacity factors by technology. Offshore wind outlook 2019. Paris:
[31] Tenhumberg N, Büker K. Ecological and economic evaluation of hydrogen IEA; 2018.
production by different water electrolysis technologies. Chem Ing Tech 2020;92 [49] Satyapal S. U.S. Department of Energy Hydrogen and Fuel Cell Technology
(10):1586–95. Perspectives. United States; 2019.
[32] Nicita A, Maggio G, Andaloro APF, Squadrito G. Green hydrogen as feedstock:
Financial analysis of a photovoltaic-powered electrolysis plant. Int J Hydrogen
Energy 2020;45(20):11395–408.

13