1 s2.0 S0360319923015410 Main

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 2 7 8 7 2 e2 7 8 9 1
Available online at www.sciencedirect.com
ScienceDirect
journal homepage: www.elsevier.com/locate/he
Utilisation of waste heat from PEM electrolysers e

Unlocking local optimisation
Els van der Roest a,b,*, Ron Bol c, Theo Fens d, Ad van Wijk a,e
a
KWR Water Research Institute, Groningenhaven 7, 3430 BB, Nieuwegein, the Netherlands
b
Faculty of Civil Engineering and Geosciences, Delft University of Technology, Stevinweg 1, 2638 CN, Delft, the
Netherlands
c
Allied Waters, Groningenhaven 7, 3430BB, Nieuwegein, the Netherlands
d
Formerly Delft University of Technology, Brasemdaal 95, 2553NC, Den Haag, NL, the Netherlands
e
Faculty of Mechanical, Maritime and Materials Engineering, Delft University of Technology, Leeghwaterstraat 39,
2638 CB Delft, the Netherlands
highlights graphical abstract
Redundant designs for utilisation Local use of waste heat Electrolyser CO₂-reduction LCOE
62°C 57°C 54°C efficiency tonne/MWh €/MWh
External heat
of waste heat from electrolysers PEM
electrolyser stack
Heat
exchanger
Heat
exchanger
user
(2,220 MWh/y)
% heat delivered
57°C 90% 0.28 8.9

are presented. Electrolyser
Hydrogen
300 tonnes/year
Techno-economic analysis for a Local use of waste heat with heat pump
62°C 57°C 100°C

direct heat consumer and district PEM
electrolyser stack
Heat
exchanger
Heat pump
External heat
user
(2,890 MWh/y) 90% 0.18 36.9
heating grid. Electrolyser

57°C
Hydrogen
300 tonnes/year
Increased electrolyser system effi- Low temperature district heating network (DHN)
62°C 57°C 54°C 44-50°C
ciency of 14e15% by waste heat PEM Heat
exchanger
Heat
District
heating
DHN
(2,365
electrolyser stack exchanger
buffer MWh/y) 91% 0.08 8.4
utilisation. Electrolyser
57°C
Hydrogen
The feasibility of heat utilisation is

300 tonnes/year
very sensitive to the transport

distance.
LCOE of the electrolyser heat is
within the range of other indus-
trial heat sources.
article info abstract
Article history: Recovery of heat from electrolysers is potentially interesting to increase the total system
Received 30 November 2022 efficiency, reduce CO2 emissions, and increase the economic feasibility of both hydrogen
Received in revised form and heat production. This study examines different designs for the utilisation of (waste)
19 March 2023 heat from a 2.5 MWel polymer electrolyte membrane (PEM) electrolyser. Redundancy is
Accepted 24 March 2023 important in the design, to ensure safe operation regardless of the heat demand of the heat
Available online 21 April 2023 consumer. We analysed cases with local heat consumption (with/without a heat pump)
and coupling with a district heating network (DHN). Overall, 14e15% of the electricity input
Keywords: to the stack can be utilised by a heat consumer, increasing the total system efficiency to
(Green) hydrogen 90% (HHV) with CO2-savings of 0.08 (DHN)-0.28 (direct use) tonne CO2/MWhheat, used. We
Electrolysis performed a first-order techno-economic analysis showing that the levelized costs of the
* Corresponding author. KWR Water Research Institute, Groningenhaven 7, 3430 BB Nieuwegein, the Netherlands.
E-mail addresses: [email protected], [email protected] (E. van der Roest).
https://doi.org/10.1016/j.ijhydene.2023.03.374
0360-3199/© 2023 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications LLC. This is an open access article under the CC BY
license (http://creativecommons.org/licenses/by/4.0/).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 2 7 8 7 2 e2 7 8 9 1 27873
Waste heat electrolyser heat (8.4e36.9 V/MWh) fall within the range of other industrial heat sources
System integration and below lower-temperature heat sources.
District heating © 2023 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications
LLC. This is an open access article under the CC BY license (http://creativecommons.org/
licenses/by/4.0/).
Nomenclature GW Gigawatt
kWh Kilowatt hour
BOP Balance of Plant
LCOE Levelized Cost Of Energy
CAPEXi Capital expenditures for a unit i (V/unit)
L Lifetime (years)
cCO2 CO2 price (V/tonne)
LCtotal Total levelized yearly costs for the heat recovery
celectricity Cost of electricity (V/kWh)
installation (V/year)
cgas Cost of natural gas (V/kWh)
LCi Levelized yearly costs for a specific system
CMargin Margin between potential savings and costs for
component i (V/unit/year)
heat (V/year)
LHV Lower Heating Value
CMargin with CO2 Margin between potential savings and costs
Loadelectrolyser Fraction of the full load of the electrolyser
for heat with CO2 pricing taken into account
MWh Megawatt hour
(V/year)
Nu Number of units
Csavings Costs savings at the heat consumer (V/year)
OM operation and maintenance costs as a percentage
CO2;reduction;heat CO2 emission reduction potential of
of the CAPEX costs (%)
replacing a fossil energy source with waste
PEM Polymer Electrolyte Membrane
heat (kg)
PHE Heat exchanger capacity (kWth)
CO2 savings Saved costs by CO2-pricing of CO2 emission
PPA Power Purchase Agreement
reduction (V/year)
Qproduced Total amount of heat produced (Watt)
COPaverage;heat pump Average coefficient of performance of a
r Discount rate (fraction)
heat pump ()
TJ Terajoule
DHN District Heating Network
Uoperating Operating voltage of the electrolysis cell (Volt)
EU European Union
Uthermoneutral Thermoneutral voltage (Volt)
Ecost Electricity costs (V/unit/year)
EElec Electricity consumption of the electrolyser (kWh) Greek letters
Eelectrolyser stack;eff Stack efficiency of the electrolyser (%) a Capital recovery factor
Eheat;used Heat from the electrolyser used by a consumer h Efficiency (%)
(kWhth)
Subscripts and superscripts
Epump Pumping energy (kWh)
el Electricity
H2;produced Amount of hydrogen produced (kg)
EL Electrolyser
HHV Higher Heating Value
H2 Hydrogen
HT-ATES High-Temperature Aquifer Thermal Energy
Nt Time steps in hours
Storage
th ¼ Thermal
I Cell current (ampere)
The European Commission has raised the ambition of

Introduction renewable hydrogen production with the REPowerEU action
plan from 5.6 Mton to 10 Mton domestic EU production and 10
Our future energy system needs to be reliable, affordable and Mton of hydrogen imports [5]. Production of hydrogen can
clean. To reach that goal, an integrated energy system needs take place both at local hydrogen clusters (‘Hydrogen valleys’)
to be created that combines smart combinations of technol- as well as at or near large RES production sites in less
ogies so energy supply and demand are matched in both time inhabited areas (oceans, deserts) [6]. At these remote locations
and space. In this future integrated energy system, green with high wind speeds or high solar irradiation, hydrogen can
hydrogen is increasingly seen as an essential energy carrier be produced, if necessary converted, and transported to areas
[1,2]. Hydrogen is a feedstock for industry as well as a zero- with high demand for either hydrogen or electricity. To con-
carbon energy carrier, that can transport and store renew- nect areas of supply (i.e. Africa, Iceland) with areas of
able energy cost-efficiently and de-carbonize energy use in hydrogen demand (i.e. Europe) [7,8] infrastructure has to be
industry, transport and buildings [2e4]. As (green) hydrogen is installed, which is reflected in the plans for a European
an energy carrier, and not an energy source, it needs to be hydrogen backbone [9,10].
produced first by water electrolysis with renewable electricity.
27874 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 2 7 8 7 2 e2 7 8 9 1
In this paper, we focus on local hydrogen clusters, as these waste heat water. This means that the waste heat it not used
clusters give more possibilities for the integration of hydrogen to its full potential. Bilbao states that when it would be
production with other sectors and services. Because the possible to utilise the full potential, it could lead to a total
electrolysis process is not 100% efficient, (waste) heat is pro- increase in efficiency of 13%, comparable to Buttler &
duced as a by-product and could be utilised by other sectors. Spliethoff [13], but they do not mention how this heat could be
The new generation electrolysers have a system efficiency of used [21]. A modelling and experimental study on an alkaline
74e79% [2,11] and a balance of stack efficiency of 77e80% electrolyser of 46.5 kW showed that the efficiency of the
(higher heating value e HHV) [12]. For most of the balance of electrolyzer system can be increased to >90% if heat recovery
plant processes (i.e. electricity transformation or demineral- is included [22]. Frank et al. [23] notice an 18% increase in ef-
ized water production) no heat can be recovered, so the ficiency if all waste heat from their 1 MWel electrolyser would
technical potential of heat recovery can only be based on the be used. Huang et al. [24] consider a model predictive control
balance of stack efficiency. Conventionally, the heat (at strategy including the waste heat of the electrolyser and
50e80 C) from the stacks is dissipated by dry coolers on top of conclude that it leads to more cross-sectoral flexibility of their
the electrolyser containers. When electrolysers are installed system, but does not quantify the amount of heat recovered.
at short distances of heat demand, the heat produced by the Hückebrink & Bertsch introduce a concept for a neighbour-
electrolyser can potentially be used resulting in a higher hood where both a fuel cell and electrolyser are installed in a
overall system efficiency. Eventually, the utilisation of waste building and their heat is used, reducing the need for a heat
heat can lead to a reduction in CO2 emissions, when the pump [25]. No quantification of the amount of heat from the
electrolyser heat is exchanged with heat from natural gas or € hm et al. [26]
fuel cell and electrolyser is given in this study. Bo
other fossil energy sources. Besides CO2 emission reduction, it give a conceptual overview of the use of electrolyser waste
will also reduce the dependence on fossil sources and could heat from low-temperature and high-temperature electrolysis
lower the hydrogen production cost. in district heating systems. Based on literature research and
Utilising waste heat from electrolyser installations is expert consultation, they conclude that there is possible
mentioned by Buttler & Spliethoff as a way of increasing the synergy between power-to-hydrogen and district heating
system (stack þ balance of plant) efficiency from 75-80% to systems and a significant potential for electrolyser waste heat
86e95% (HHV based) [13]. They mention three concrete pro- of temperature levels below 100 C. They state that there is a
jects where heat integration was part of the activities. Firstly, need for more quantitative research. Overall, we conclude
the BioCat project deployed a 1 MW installation to produce that both the scientific community and industry recognize the
hydrogen via electrolysis and converted it to methane with potential of electrolyser waste heat, yet both publications and
the help of biological methanation. The intention was to reuse practical examples are scarce. Moreover, there is a lack of
both the heat from the electrolyser and methanation process detailed system design and analysis for the utilisation of
in the waste water treatment plant or a renewable heat grid. waste heat from electrolysis.
Eventually, the utilisation of heat from the electrolyser was
not realised in the project (p. 6 final report) [14]. Secondly, Research goal
Stromlückenfuller is a project where waste heat from a
200 kW electrolyser is used in a heating network [15]. A similar Based on the scarce availability of literature and concrete
project with a 1 MW electrolyser and reuse in a district heating examples of waste heat utilisation from electrolysis, our
grid is Green Hydrogen Esslingen [16]. Lastly, RWE uses heat research question is:
from electrolysis (150 kW) in a gas pressure regulation station What is the potential of waste heat from electrolysers and
[17], but no further publications are found on how these sys- how could it be utilised?
tems work and how much heat is recovered. It is thus hard to To answer this question, we will elaborate on different
find concrete examples of the realization of waste heat re- designs for the utilisation of electrolyser waste heat based on
covery, but new plans for heat utilisation are created. In a 2.5 MWel Polymer Electrolyse Membrane electrolysis (PEM)
Hamburg, a 100 MW electrolysis system is proposed with electrolyser (stack size) in Nieuwegein, the Netherlands. The
waste heat utilisation in the district heating network and electrolyser is part of the H-Flex project, which investigates
thermal treatment of municipal waste, that should be opera- multiple ways in which the electrolyser can be utilised next
tional in 2025 [18]. The 2050 scenario for the district heating to hydrogen production, by offering flexibility to the elec-
network of Aalborg, the utilisation waste heat from a 330 MW tricity grid as well as waste heat utilisation [27]. The theo-
electrolyser is mentioned as an additional heat source [19]. retical heat utilisation potential can be 500 kWth of heat at
The Institute of Process Technology has worked out a design 100% load at the start of the lifetime, based on 80% efficiency
for a 1 GW electrolysis plant, which should be heat integration (HHV) of the electrolyser stack. How heat can be utilised will
ready [20]. differ per use case. We will work out three different use cases
The scientific literature on heat utilisation of electrolysers that represent possible situations for local system integra-
is limited. Bilbao [21] worked out a case study in Chile for an tion. In an ideal case, there is a heat consumer that uses heat
alkaline electrolyser, with as goal to provide the electrolyser at the output temperature level of the electrolyser. As a
with pre-heated water to have a higher overall efficiency. The second case, we include a heat pump for higher-temperature
overall efficiency increase was marginal (<1%). The reason for heat. The third use case is the delivery of heat to a local
the small overall efficiency increase is not the amount of district heating system. We will perform a first-order techno-
available waste heat, but the relatively small amount of water economic analysis on these three use cases including cal-
consumption, which is 30 times lower than the mass flow of culations on how much heat can be utilised and what the
combined system efficiency will be. Furthermore, the po- that is separated from the hydrogen/water stream by the gas-
tential CO2 reduction is calculated when electrolyser heat liquid separator. The oxygen/water stream flows to an oxygen
replaces heat from fossil sources. Finally, the costs of heat separation tank which is combined with the deionized (ultra-
recovery and transportation are calculated to get a first idea pure) water recirculation stream, with a heat exchanger on
of the feasibility of heat recovery in general and specifically the stream between the oxygen separation tank and the
for these three cases. recirculation pump. Excess heat is removed to limit the inlet
temperature to the stack if the condensing heat from the ox-
ygen flow exceeds the heat needed to raise the temperature of
Methodology the demineralized water inlet. Temperature levels for PEM
electrolysers are mostly reported in the 50e80 C range
Design of a PEM electrolyser with heat utilisation [13,20,29]. We have considered here an average working
temperature of the electrolyser of 65 C, and a cooling circuit
In this publication, we focus on a 2.5 MW PEM electrolyser with a maximum inlet temperature of 57 C and a DT of 3 C
installation in Nieuwegein [27], although our method could over the heat exchangers [29].
potentially be applied to other electrolyser capacities as well. This is an efficient design where heat is extracted from
In water electrolysis, water is converted via an electro- both the oxygen and (indirectly) the hydrogen stream and
chemical process with two electrodes into hydrogen and ox- directly used to heat the deionized water stream to the elec-
ygen. The chemical reactions in a PEM electrolyser are given in trolyser. If the temperature of the stack is known, then one
Eqs (1)e(3), with a change in enthalpy DH of 285.84 kJ/mol, can calculate the amount of heat necessary to heat the
which is the energy required to drive the reaction. incoming water based on the hydrogen production. According
to the thesis of Tiktak [29], this amounts to approximately
Cathode : 2Hþ þ 2e /H2 (1)
5e8% of the total available heat. The excess heat is removed
by a dry cooler. In case the excess heat would be used by a
1
Anode : H2 O / O2 þ 2 Hþ þ 2e (2) third party, the design has to be adapted to make this possible
2
with an extra tie-in on the cooling system. Choosing for and
1 extra tie-in on the cooling system instead of replacing the dry
Overall : H2 OðlÞ / H2 ðgÞ þ O2ðgÞ DH ¼ 285:84 kJ = mol (3) cooler makes the system redundant.
2
Heat sources in an electrolyser Variance in heat production

Heat is produced in the electrolyser stack due to the irre- Two factors lead to variation in heat production; the efficiency
versibility of the chemical reactions happening and the ohmic load curve of the electrolyser stack and the degradation pro-
resistance of the cell. The electrolysis reaction can occur cess. Firstly, the efficiency is negatively correlated with the
without consumption or production of heat, at the so-called load, therefore we use an efficiency curve (Eq (5), visualised in
thermoneutral voltage. However, at this potential, the reac- Fig. 2) that varies with the load based on empirical statistics
tion rate is too low, so there is almost no hydrogen production. analysis [30], but with efficiency values adapted to recent de-
Therefore, an overvoltage is applied to the cells, which in- velopments. Eq. (5) gives an electricity consumption of about
creases the reaction rate and makes the electrolysis reaction 49.25 kWh/kg at 75% load or an expected balance of stack ef-
exothermic, thus heat is released. The total amount of heat ficiency of 80%which is the expected stack efficiency of the
produced (Qproduced , in Watt) is the cell current (I in ampere) PEM electrolysers that are currently available [11,31].
times the difference between the operating voltage (Uoperating in
HHVH2
Volt) and the thermoneutral voltage (Uthermoneutral , in Volt) is Eelectrolyser stack;eff ð>10% loadÞ¼ $100
7:8455 kWh
kg
* Loadelectrolyser þ
43:409kWh
kg
given in Eq (4).
(5)

Qproduced ¼ I$ Uoperating Uthermoneutral (4)
With Eelectrolyser stack;eff the electrolyser stack efficiency in %,
In some cases, the released heat is in balance with the heat Loadelectrolyser represented as a fraction of full load (i.e. 0.8) and
consumption for the evaporation of water in the stack, but the HHVH2 being 39.4 kWh/kg. Eq. (5) is only valid above 10%
otherwise, this heat needs to be removed from the electrolysis load. Below 10% load, the efficiency sharply decreases. In this
stack to maintain the cell temperature and prevent over- study, we will only consider electrolyser capacities above 10%
heating. Heat can be extracted from the stack from three load.
different sources in a PEM electrolyser [28,29]; from the oxygen Secondly, the ageing and degradation of the electrolyser
stream, the hydrogen stream and the water recirculation loop. stack lead to a lower hydrogen production efficiency. Mainly
Both the oxygen and hydrogen flow contain some water due to the degradation of the stacks, the resistance over the
vapour after leaving the stack from which heat can be membranes increases and therefore requires a higher over-
extracted. potential over time. This linear voltage degradation can result
Based on the design of Mancera et al. [28], the one GigaWatt in a cell voltage increase. With an 80,000-h lifetime, this would
electrolyser project [20] and information from electrolyser result in a voltage degradation of 1.9 Ve2.06e2.22 V [13].
manufacturers, Fig. 1 shows a simplified cooling system of an Therefore, the amount of heat production will increase over
electrolyser. Heat is extracted from the recirculating water the lifetime of the project if the hydrogen production rate is
Fig. 1 e Cooling system of an electrolyser stack. In orange/brown the oxygen flow is shown, in (dark)green the hydrogen
flow and in dark blue the water flow. The cooling water flow is shown in red and blue. (For interpretation of the references to
color in this figure legend, the reader is referred to the Web version of this article.)
Fig. 2 e Electrolyser balance of stack efficiency in relation to the load, based on Eq (5), [11,30,31].
kept constant. An evaluation of several electrolyser projects in Let's illustrate this with an example of electrolysis (electro-
the EU [32] shows that these PEM electrolysers have a degra- chemical conversion, so no burning process is involved), where
dation of 0.12%/1000 h, which is the FCHJU target for 2030. The the LHV would lead to an overestimation of the available heat
final reported value for PEM degradation is 0.19% per 1000 h, from electrolysis. If we would assume 80% electrolyser stack
which means after a year of full production (8000 h) 1.5% extra efficiency on HHV, this amounts to an electricity consumption
energy consumption and 15% extra energy consumption after of 49.25 kWh/kg hydrogen by the electrolyser. The LHV value of
10 years [32]. Suppliers on the other hand report <1.0% effi- hydrogen is 33.3 kWh/kg or 121 MJ. The LHV stack efficiency
ciency degradation [12]. We will adopt the lower-end values would then be 33.3/49.25 * 100 ¼ 68.2%. Using the LHV for the
for newer generation electrolysers, thus assuming a 1.0% calculation of the heat recovery potential of electrolysis would
annual efficiency degradation which will lead to a 10% in- lead to a perceived heat recovery of 31.8%. While at HHV, a
crease in energy consumption after 10 years. stack efficiency of 80% means there is a theoretical heat re-
covery potential of 20%. By using the LHV, one would thus
HHV vs LHV overestimate the heat recovery potential by more than 50%. So,
calculating with the HHV is required to arrive at a univocal
The chemical energy content of a fuel is expressed as the HHV energy balance in the electrolyser/fuel cell application. It pro-
(higher heating value). For hydrogen, this is 39.4 kWh/kg or vides more information on the actual energy content of
141.8 MJ/kg. In addition, there is also a lower heating value hydrogen, as well as how much waste heat is available. In this
(LHV, 33.3 kWh/kg or 120 MJ/kg), which is only relevant when a publication, we will therefore consistently use the HHV.
substance is burned and no heat is recovered from flue gases.
When the latent heat from flue gases is recovered, there is no Case studies
burning process involved, or hydrogen is used as feedstock, it
is misleading to use the LHV as it leads to an overestimation of The waste heat utilisation of the electrolyser was analysed by
the waste heat potential. three relevant use cases. For all cases, the electrolyser details as
presented in Table 1, based on the 2.5 MW (stack size) instal- available electrolyser heat. In the case of Nieuwegein, this is
lation in Nieuwegein. The hydrogen production pattern was an industrial-scale laundry washing company, but it could be
based on the yearly production target, scaled down to an hourly a different user as well. We considered two options, either
production target. The electrolyser is stimulated to run at 75% the heat was directly utilised at the right temperature, or the
of its capacity, on local solar PV as much as possible (based on temperature was raised by a (high-temperature) heat pump,
weather data) or during hours of low electricity prices (based on see Fig. 3. High-temperature heat pumps with temperature
day-ahead prices). In case the production capacity stayed levels of up to 100 C are proven technology [34,35]. In Table 2
behind the target (based on the sum of the production), the the chosen parameters for case 1 are shown. The maximum
capacity target was increased to 90%. In hours where no heat production of the 2.5 MW electrolyser (at 100% load)
renewable capacity is available, the capacity was limited to would be 580 kWth up to 640 kWth after 10 years. Yet, most of
75%. In our analysis, we only considered the stack size for heat the time the electrolyser will work at a lower load (i.e. 75%)
recovery, as no heat will be recovered from the auxiliary and not all heat can be recovered, hence the heat exchanger
equipment. The efficiency of the total electrolysis installation is size was chosen to be 400 kWth. The heat consumer has a
thus lower. The electrolyser worked on average at a 75% load larger heat load than the electrolyser offers and has a heat
and 80% efficiency (HHV), so the theoretical waste heat poten- buffer available which means the electrolyser heat can be
tial of the electrolyser stack would be 20% of the energy input. stored there outside of the working hours of the company.
The technical potential for heat recovery was set at 80%, which The heat load of the consumer is therefore not modelled in
means 16% of the electrolyser waste heat was recovered. Tiktok detail, but it is assumed that in general the electrolyser will
found a recovery percentage of 92% [29] with internal cooling in be able to deliver its heat to the consumer. Yet, the buffer
the stack inside the bipolar plates. Our design has cooling could be full at a certain point, therefore it is not likely that
outside of the stack, with multiple heat exchangers (see Fig. 1) all electrolyser heat can always be utilised. We have
and more heat losses, therefore we have chosen a value of 80%. assumed that 80% of the available heat can be used by a heat
Over time, the heat potential increased (see 2.1.2) which at 75% consumer.
load (1.875 kW) meant 300 kWth of heat in the first year of
operation which increased to 330 kWth in year 10. Heat losses Case 2 - low-temperature district heating network
during heat transport are based on calculations of a DN80 pipe The electrolyser heat could also be used in district heating
(90/162 mm) with insulation (class 1) based on heat loss norms systems. The temperature level fits best for low-temperature
for heat pipes [33]. For pump energy, a pump efficiency of 60% is heating systems (4th or 5th Generation [36]), with a tempera-
assumed. ture level of 40e50 C/25 C (supply/return temperature). In an
earlier study, we analysed a case of 2000 houses in a neigh-
Case 1 - local use of heat bourhood at a distance of about 3 km from the electrolyser. In
If the electrolyser is placed in an area with more industrial the ‘Power-to-X’ scenario, the heating system consisted of a
applications, there may be a direct consumer for the district heating network (DHN), heat extraction from surface
water with a heat pump, seasonal heat storage in a high-
temperature aquifer thermal energy storage (HT-ATES) sys-
Table 1 e Electrolyser parameters. tem and a day buffer [37]. Space heating was supplied directly
via the DHN, for domestic hot water (tap water) additional
Electrolyser parameters
booster heat pumps are installed in buildings or houses to
Size of electrolyser stack 2.5 MWel provide safe tap water of at least 60 C to households. Heat
Auxiliary equipment e Energy use is ca. 5% of stack
demand was varied based on house type and hourly weather
outside of scope capacity
data via the degree-day method [38]. We have considered the
Production target 300 tonnes/year (capacity factor
0.71) use of waste heat from an electrolyser there as well, but with a
Hydrogen production 80% at 75% load, see further Eq (5) strict assumption that the waste heat could only be delivered
efficiency (HHV) to the DHN if it would fulfil the total heat demand in that hour.
Minimum load 10% Then, only 1.5% of the waste heat could be utilised [37]. In the
Waste heat recovery 80% current publication, we took a more flexible approach. Instead
efficiency [29]
of delivering heat directly to the DNH, the electrolyser heat
Available heat at 75% load Year 1: 300 kWth
Year 10: 330 kWth
was delivered to the day buffer of the DHN, see Fig. 3 and Table
Heat for water flow to stack 0.571 kWh/kg H2 produced 3 for details. The buffer adds the necessary flexibility so the
(DT ¼ 55 C) electrolyser heat could be combined with heat from other
Electrolyser working 65 C sources. We assumed that the electrolyser heat can be added
temperature without limitations to the day buffer of the DHN. The day
Full load hours of the 6250
buffer of the DHN was estimated at 600 m3 (21 MWhth) to
electrolyser
accommodate 12 h of average heat demand (1.750 kWhth),
Cooling water inlet 57 C
temperature [29] while the electrolyser delivers a maximum of 400 kWhth of
Cooling water outlet 62 C heat to the buffer in 1 h. In case the buffer is full and there is
temperature [29] low heat demand, heat could be stored in the HT-ATES sys-
Loss over heat exchanger 3 C tem. Moreover, in the calculation, we took into account the
Annual efficiency 1.0% ageing effects of the electrolyser leading to an increase in heat
degradation [13,20,32]
availability over time.
Fig. 3 e Schematic overview of the different cases; 1a e Local use of waste heat at low temperature, 1b e Local use of waste
heat at high temperature, 2 e Delivery of waste heat to a low-temperature district heating network. The yellow arrows show
which electricity demands are taken into consideration (pumping energy is included). Heat losses during heat transport are
based on calculations of a DN80 pipe (90/162 mm) with insulation (class 1) based on heat loss norms for heat pipes [33]. (For
interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Efficiency & energy balance PNt PNt

H2;produced;i $HHVH2 þ Eheat;used;i
hEL ¼ i
PNt
i
$100% (6)
i EElec;i
The amount of heat utilisation was calculated on an hourly
basis with the Power-to-X techno-economic simulation model
based on the Power-to-H3 concept introduced in earlier pub-
lications [37,43]. Based on varying hourly demand profiles,
Table 3 e Parameters for heat delivery to a district heating
hourly weather data, and data about conversion and storage
network.
technologies (batteries, power-to-heat, power-to-hydrogen,
long-term heat storage and hydrogen storage), an hourly en- District Heating Network
ergy balance was calculated for multiple years. Because the Preferred temperature 45e60 C
electrolyser degrades over time while the hydrogen produc- level - low
Total neighbourhood 55 TJ/y
tion target is constant, the electricity demand increased over
heat demand
time, as well as the heat production. Both a yearly and 10-year
Distance between 3 km
average efficiency of the electrolyser stack (hEL in %) were production and use
calculated according to Eq (6). Distance between waste 200 m
heat production and
day-buffer/HT-ATES
Size heat exchanger 400 kWth
between electrolyser
Table 2 e Case 1 - Direct heat consumer parameters. and day buffer
Local consumer Heat source Surface water þ heat storage (HT-
parameters ATES)
Average COP large-scale 5
Preferred temperature level e low (1a) 54 C
heat pump [39,40]
Preferred temperature level e high (1b) 100 C
Day buffer charging/ 90%/100%
Distance between production and use 200 m
discharging efficiency
Heat exchanger capacity 400 kWth
[41]
% of overlapping working hours of 80%
Day buffer heat loss per 2%
electrolyser and heat consumer
weeka [42]
(including buffer)
a
High-temperature heat pump capacity (1b) 520 kWth Data are given for 500 m3 with 2.1% heat loss, as we need a
Average COP heat pump (1b, see SI.1) 4.3 smaller tank here, we increased the heat loss to 3%.
With H2;produced the amount of hydrogen produced in a year calculated by Eq (10) by summing over the number of units Nu .
(when i.e. I ¼ 1 and the number of time steps in hours The calculation was made from the perspective of the elec-
Nt ¼ 8760) or over the total period (i ¼ 1 and n ¼ 87,600) in kg, trolyser owner delivering heat to a heat consumer. General
HHVH2 the higher heating value of hydrogen (39.4 kWh/kg), cost parameters used in the calculations are given in Table 5.
Eheat;used the amount of useful heat from the electrolyser in XNu
P LCtotal ¼ LCi with (10)
kWh over the period i to Nt , and ni EElec the electricity con- i
sumption of the electrolyser in kWh over the period i to Nt in

hours. LCi ¼ a$CAPEXi þ OMi þ Ecosti (11)
Where LCi represents the annual levelized costs for a certain

CO2 emission reduction
system component in (V/unit/year). The CAPEXi (V/unit) are
the capital expenditures for a particular system component i
If the heat from the electrolyser replaces heat from fossil
such as the pumps, the pipes or the tie-in, see Table 6 for an
sources, there is a CO2 reduction potential. The amount of CO2
overview of components. More information on the specific
reduction depends on the use case. For case 1, the local use of
components and their costs is given in SI.1. The calculations
heat by a third party, we assume that this heat was otherwise
included the necessary costs for heat recovery at the electro-
produced by natural gas. For the gas boilers, we assumed an
lyser, transportation costs and control technology. It did not
HHV efficiency of 90% (hboiler ) for the industrial boiler. The CO2
include an installation on the consumer side as it is not known
emission reduction potential in kg for heat CO2;reduction;heat is
how the heat will exactly be integrated into the process of the
calculated by Eq (7) for case 1a, and by Eq (8) for case 1 b,
heat consumer.
including the heat pump electricity consumption. We
The OMi (V/unit/year) represent the operational expendi-
assumed that the electricity for the heat pump is not certified
tures for a particular system component, which are a per-
renewable electricity, and therefore a CO2 emission factor is
centage of the CAPEX costs. Ecosti (V/unit/year) are the
added. The reference situation for case 2 (DHN) is a heat
pump. For the heat pumps, we calculated the maximum po-
tential for CO2 emission reduction by using the CO2 emission
factor for the average grid electricity mix in the Netherlands, Table 5 e General cost parameters.
see Eq. (9). In all cases, the pumping energy necessary for Parameter Value
transporting the heat to the heat consumer Epump (in kWh) is OM 2% of CAPEX
taken into account. Discount rate (r) 5%
Installation factor pipes 1.2
Eheat;used Installation factor for pump, heat 1.4
CO2;reduction;heat;case 1a ¼ CO2;emissionfactor;gas $ Epump (7)
hboiler exchanger, tie in & electronics
Electricity costs (for pumps/heat 0.0834 V/kWh
Eheat;used pumps)a (celectricity )
CO2;reduction;heat;case 1b ¼ CO2;emissionfactor;gas $ Gas price industrial usersa (cgas ) 0.0341 V/kWh
hboiler
CO2 priceb ðcCO2 Þ 60 V/tonne
Eheat;used
þ CO2;emissionfactor;electricity $ Epump (8) a
COPaverage;heat pump Price of gas and electricity for industrial users in the 10e1000 TJ
(gas) or 2000e20,000 MWh (electricity) range, including taxes,
excluding VAT, average price 2016e2020. Prices are converted
Eheat;used
CO2;reduction;heat;case2 ¼CO2;emissionfactor;electricity $ Epump from GJ to kWh [45].
COPaverage;heatpump b
The CO2 price is based on the middle value of the carbon pricing
(9) benchmark of the OECD [46], which is a conservative number
compared to recent developments [47].
With COPaverage;heatpump the average COP of the heat pump
over the period analysed.
The emission factors (CO2;emissionfactor ) in kg/kWh are given in
Table 4. Table 6 e Individual component costs parameters. In SI.1,
more information is given on the cost calculations for
Cost calculation specific system components.
Component CAPEX Lifetime
The levelized yearly costs for the heat recovery installation (L, in years)
(LCtotal in V/year) including delivery at the consumer were Tie-in electrolyser 10 kV 15
Electronics, controls & 40 kV 10
monitoring
Table 4 e CO2 emission factors for natural gas and Pumps see SI.1 15
renewable electricity, from a CO2 emission factors Pipesa 230 V/m (2019) 40
database for the Netherlands [44]. pffiffiffiffiffiffiffiffi
Heat exchanger 1500* PHE (2017) 20
Well-to-wheel Heat pump (including 600 V/kWth 20
CO2 emissions installation)b e case 1 b
a
Natural gas e case 1 & case 2 2.085 kg/Nm3 or Based on the Polytherm pricelist (2019) [48].
b
25.7 g/kWh (HHV) The investment is including installation, approx. 50% of costs
Grid electricity (grid mix The Netherlands) 42.7 g/kWh [49].
electricity costs for a system component i., such as the elec- CO2;reduction;heat;casej
tricity costs for pumps. The capital recovery factor (a, no unit) CO2 savings;casej ¼ cCO2 $ (17)
kg
1000 tonne
is a fraction of the total CAPEX cost. It represents a constant
yearly value of depreciation and is calculated based on the The CO2 cost savings are taken into account in the LCOE as
project lifetime (L, in years) and the discount rate (r, as a a reduction of the yearly costs according to Eq (18).
fraction of 1) (Eq (12)). In the cost calculations, we allocated the
investments of the heat recovery installation completely to LCtotal CO2 savings
LCOEheat ¼ (18)
the costs of waste heat. The costs of the electrolyser were Eheat;used
however excluded from the costs calculation of the waste heat In the margin, they are added to the other cost savings (Eq
as the investments for the electrolyser installation are done (19)).
regardless of the waste heat consumption. Yet, if there is a
positive business case for waste heat utilisation, this could CMargin with CO2 ;casej ¼Csavings;casej þ CO2 savings;casej LCtotal;casej (19)
lead to a lower hydrogen price.
Sensitivity analysis
r
a¼ (12)
1 ð1 þ rÞL To assess the effect of changes in variables that are still uncer-
Based on the yearly costs, the levelized costs of heat tain, or from which a large effect is expected, a local sensitivity
(LCOEheat ) in V/MWh were calculated according to Eq (13), with analysis was performed. In Table 7 the parameters chosen for
Eheat;used the amount of heat utilised by the heat consumer in the sensitivity analysis are shown, including the ranges over
MWhth/year. which the parameters were varied and in which cases they are
applied. The annual efficiency degradation, tie-in costs and
LCtotal costs for electronics are relevant and these numbers are hard to
LCOEheat ¼ (13)
Eheat;used verify as there are no installations yet in place. Therefore, we
To know if the business case is positive, we calculated the assessed their influence on the LCOE to know how a change in
cost savings for the consumer. Then, we calculated the value would affect the LCOE. The overlapping working hours,
margin between yearly costs and savings. If the margin is discount rate, heat pump temperature, the distance between
positive, it means there is a possibility for a business case as production and use and the electricity price are likely to have a
the saved costs for the consumer are higher than the costs for significant influence on the LCOE. Moreover, these values are
heat delivery. The costs savings for the consumer (Csavings ) in likely to vary for other use cases, so the sensitivity analysis gives
V/year were calculated in euros by using heat from the elec- more general insights as well. High-temperature heat pumps
trolyser instead of gas (case 1) or electricity (case 2). The costs with temperature levels above 100 C are not yet standard
savings are based on the average amount of Eheat;used over the products, but there are suppliers available with a technology-
total project lifetime (in kWh/year), see Eq. (14) and (15). readiness level (TLR) of 7e9 for temperatures up to 200 C [50].
Here we have chosen a maximum of 150 C, so as to not over-
Eheat;used estimate the possible temperature lift at the heat consumer side.
Csavings;case1 ¼ cgas $ (14)
hboiler The gas price is not taken into account in the LCOE but only
in the saved costs by the heat consumer, therefore, we will
Eheat;used assess this variable for the margin. As gas prices have recently
Csavings;case2 ¼ celectricity $ (15)
COPaverage;heat pump been shown to be highly sensitive to geopolitical circum-
With cgas the costs of natural gas in V/kWh, hboiler the gas stances, we have taken extremely high prices into account.
boiler efficiency (presented as a decimal value) of the heat The range for the CO2 price is chosen to be 0 (no CO2 pricing) to
consumer and celectricity the costs of electricity in V/kWh. In case
2 (DHN), the heat would otherwise have been produced by an
industrial-scale heat pump. In this case, the amount of heat Table 7 e Ranges for sensitivity analysis.
from the electrolyser was divided by the average COP of the
Number Range Case
heat pump COPaverage;heat pump over the same time period.
Annual efficiency 1.0% 0.5e2% 1,2
Finally, the margin per case (CMargin;casej in V/year) between
degradation
potential savings and costs for the heat, recovery installation
electrolyser [13,20,32]
was calculated as an outcome according to Eq. (16). Overlapping working 80% 20e100% 1,2
hours
CMargin;casej ¼ Csavings;casej LCtotal;casej (16)
Discount rate 5% 2e10% 1,2
Yet, these calculations are highly sensitive to the cost as- Tie-in costs 10.000 V ±30% 1,2
Electronics 40.000 V ±30% 1,2
sumptions of gas and electricity prices. Therefore, a sensi-
Heat pump temperature 100 C 80e150 C 1b
tivity analysis was done (see section 2.7) to investigate the
Distance between 200 m 100e5000 m 1,2
effects of energy prices on the business case. production and use/
Lastly, the effect of CO2 pricing has been taken into account storage
for every case j as optional cost savings with Eq. (17), with Electricity price 0.0834 V/kWh 50/þ500% 1,2
CO2 savings;casej the saved costs by CO2-pricing of CO2 emission Gas price 0.0341 V/kWh 25/þ500% 1
CO2 price 60 V/tonne 0-120 V/tonne 1,2
reduction in V/year. cCO2 is the CO2 price in V/tonne.
120 V/tonne based on the high range of the carbon pricing cooling system to create a redundant design independent of
benchmark of OECD [46]. the heat consumer. The electrolyser can thus either be cooled
through heat consumption by an external heat consumer or
by the dry cooler when there is no external heat consumption.
Results & discussion We choose deliberately for a tie-in on the cooling system
instead of replacing the dry cooler, as this makes the system
Design for heat recovery from a PEM electrolyser redundant. Both the hydrogen producer as well as the heat
consumer should continue their processes regardless of the
To utilise the electrolyser heat, we propose a tie-in on the delivery or demand of the other to give them the most
cooling system as shown in Fig.4. The tie-in is an add-on to the flexibility.
Fig. 4 e Redundant designs for heat recovery from a PEM electrolyser.

Table 8 shows the simulation results for all cases,
savings utilised effi-ciency demand lyser heat savings

(Electrolyser (tonne/ (MWh) (stack þ neighbour of total (tonne/
year)
starting with the reference case. The amount of available
% electro- CO2
150
160
170
180
190
200
205
210
220
225
190
heat increases over time (Fig. 7) due to the degradation of
the electrolyser stack (for specs see Table 1). The electric
demand
resistance increases and thereby the electricity consump-
heat
10%
14%
14%
13%
18%
17%
16%
18%
18%
19%
16%
tion. This extra electricity consumption is not converted to
hydrogen, but to heat instead. At the start, about 20% of the
electricity input to the stack is converted to heat. After ten
(MWh)
Case 2
years, the total electricity consumption has raised by 10%,
19,060
14,420
15,730
17,080
12,860
14,040
15,760
14,750
15,270
14,450
15,340
hood
Heat
so now 30% of the electricity input to the stack is converted
to heat, which means the heat production increased by
50%. In the reference case, this heat is not recovered
leading to an average (HHV) efficiency of 76% for the elec-
Total
heat)
92%
92%
92%
91%
91%
91%
90%
90%
90%
89%
91%
trolyser stack.
Table 8 e Results overview table. Case 1a and 1b have the same total (stack) efficiency, thus these values are only presented once.
Efficiency & energy balance
Heat
1900
2020
2130
2230
2330
2440
2530
2610
2690
2780
2365
The reference situation without heat utilisation can be
compared to three cases where the electrolyser heat is uti-
year)
CO2
430
450
480
500
510
530
560
570
590
610
525
lised, overall results are shown in Table 8. The first case has a
heat consumer that utilises the electrolyser heat directly with
an 80% overlap in electrolyser heat production and heat con-
utilised
(MWh)
& heat
pump)
sumption, thereby reducing gas consumption by a gas boiler.
Heat
2360
2490
2630
2740
2840
2950
3060
3170
3280
3380
2890
Case 1b
On average 2220 MWh/year of heat can be utilised by the heat
consumer in case 1a, which leads to a total system efficiency
of 90% (stack þ waste heat, HHV based). In the Sankey dia-
consumption production production efficiency utilised efficiency savings consumption
year) pump (MWh)

Elec-tricity
gram, all different energy flows are shown including the heat
losses and amount of unused heat (see Fig. 5) based on a ten- heat
540
580
610
630
650
680
710
730
760
780
665
year average. Pumping energy is about 7 MWh/year, rounded
off to 5 MWh and too small to be visible in the Sankey diagram.
The increase in heat production and consumption over time
(MWh) (stack þ (tonne/
and the overall efficiency are shown in Fig. 7. Eventually,

CO2
510
540
560
590
610
630
660
680
700
730
620
almost 59% of the heat that is formed in the electrolyser is
utilised by the consumer. Similarly, we find that 14% of the
Case 1 a
electricity input in the electrolyser is used as heat by a heat

Total
heat)
92%
91%
91%
91%
90%
90%
90%
89%
89%
88%
90%
consumer. Compared to the base scenario without heat uti-
lisation, the average 10-year efficiency increases by 14% from
76% to 90%. In case 1b, with a heat pump to raise the tem-
Heat
1820
1920
2020
2100
2190
2270
2360
2440
2520
2600
2220
perature of the electrolyser heat, 2890 MWh/y of heat is
available on average. The amount of heat delivered by the
electrolyser stays equal, but electricity has been used to in-
(HHV)
Base
crease the temperature, thereby increasing the DT and thus

79%
79%
78%
77%
76%
75%
75%
74%
73%
72%
76%
Reference case
the amount of heat available at the consumer. The efficiency

of the electrolysis system in case 1b is thus equal to case 1a
and therefore not shown separately.
(MWh)
Heat
2270
2400
2540
2660
2780
2910
3050
3180
3310
3450
2855
In case 2, the electrolyser heat is reused in a low-

temperature district heating grid (Fig. 6), and the heat from
the electrolyser is supplied to the heat buffer of the DHN. The
total system efficiency raises to 91% when the heat from the
(tonne)
299
300
301
301
300
300
301
300
300
300
300
electrolyser is used in a low-temperature DHN. The amount of

H2
heat provided increased over the years and accounts on

average for 16% of the total heat demand of the neighbour-
hood (2000 houses) supplied by the DHN. The variation in the
electrolyser
Electricity
(MWh)
percentage of fulfilled heat demand by the electrolyser is

14,840
15,040
15,240
15,350
15,510
15,670
15,880
16,010
16,170
16,340
15,605
caused by a difference in heat demand over the years due to

varying climate data (the years considered are 2010e2019).
About 15% of the electricity input of the electrolyser is finally
Average
used as heat in the DHN. Interesting to note is that the elec-

Year
Y10
trolyser waste heat is not used to fill the seasonal heat buffer,
Y1
Y2
Y3
Y4
Y5
Y6
Y7
Y8
Y9
but only the day buffer. Thus, the daily heat demand in
Fig. 5 e Sankey diagram for the energy balance of an electrolyser system with an industrial (local) heat consumer (1a). The
10-year average energy flows are shown.
Fig. 6 e Sankey diagram with an average yearly energy balance when electrolyser heat is provided to a DHN (case 2), 10-year
average flows are shown.
Fig. 7 e Heat production by the electrolyser and heat used by the heat consumer shown over 10 years for case a1 (local heat
consumer), including the reference case of the electrolyser (without heat recovery) and the efficiency including heat recovery
(total efficiency of stack þ waste heat utilisation). Efficiency (HHV) is shown on a secondary axis.
summer is still high enough to take up the daily heat produced When taking the saved yearly costs into account, we note
by the electrolyser. that all scenarios have a positive margin, so a possible busi-
ness case. Although in the current analysis, only a heat
CO2 reduction potential for local heat use (case 1) exchanger (or heat pump) has been taken into account at the
heat consumer, so there are possible other costs for adapta-
The current energy source of the heat consumer is natural gas. tions at the heat consumer to use the heat that have been
When part of its heat demand is fulfilled by the electrolyser neglected. Thus, a margin just above zero for case 1b, makes it
heat, on average 620 tonnes of CO2 are mitigated each year, or uncertain if a business case can be realised. For the heat de-
270,000 m3 of natural gas (0.845 kg/Nm3 [51] and a boiler effi- livery without a heat pump, the margin is almost 65 kV/year.
ciency of 90%). In more general terms, this means 0.28 tonne For case 2 (DHN) the margin is also positive (18.5 kV/year),
CO2 per MWh of waste heat used (MWhheat, used). For the case although it is a factor 3 less than for the industrial heat con-
with a heat pump (1b), the CO2 emission reduction is 525 sumer, while more heat is delivered. Our reference case for
tonnes per year or 0.18 tonne CO2/MWhheat, used. We assumed the DHN was heat produced with a heat pump that has a COP
that the electricity for the heat pump is not certified renew- of 5, and not gas boilers. Therefore, just as with the CO2 cal-
able electricity, which results in more CO2 emissions for the culations, here we see again the effect of the high COP of the
use of electricity by the heat pump than the savings by heat pump. The heat pump produces heat with only 0.2 units
switching from natural gas to electricity. Yet, the amount of of electricity for one unit of heat. Hence, the savings in elec-
saved natural gas is higher in this case, about 350.000 m3. If tricity costs in case 2 are less than the gas savings in cases 1a
the heat consumer would buy or produce renewable elec- and 1b. How the margin is divided among the stakeholders
tricity, the CO2 emission reduction would be higher than in (electrolyser owner, heat user and possibly others) will finally
case 1a. determine the effect on the heat costs for the heat consumer
For the DHN network (case 2), the CO2 emission reduction as well as the hydrogen production costs. If CO2 pricing is
is on average 190 tonnes/year or 0.08 tonne CO2/MWhheat, used, taken into account (Table 10), the LCOE decreases substan-
compared to when the electricity would have been produced tially while the margin increases. All margins are now at least
with a heat pump working on an average grid-electricity mix. 30 kV/year, pointing towards a high possibility for a business
The heat pump has an average COP of around 5, so only 1/5th case. For direct heat use (case 1a) the LCOE becomes negative.
of the heat demand is reflected in electricity demand. On the In practice, this means that a company can save more money
other hand, the CO2 emission factor for electricity is higher on CO2 costs than it would cost to replace natural gas with
than for gas (due to conversion losses). In summary, the CO2 waste heat, not even including the saved costs of the natural
reduction will be the highest when the electrolyser heat would gas itself. Upgrading the heat with a heat pump to 100 C could
be used directly without a heat pump (1a), as long as the heat also be economically feasible if CO2 costs are taken into
pump at the heat consumer is not running on green account.
electricity.
Sensitivity analysis
Economic results
We have investigated the impact of a change in certain input
The results of the economic analysis are shown in Table 9 for parameters on the LCOE, in Fig. 8 the results of the sensitivity
all three scenarios. A detailed breakdown of the costs per analysis are presented. Although in the case of Nieuwegein,
component including OM costs and energy costs is given in the distance is known, in other cases of heat utilisation the
Table S2-S5 of the supplementary information (SI). The yearly distance between the electrolyser and heat consumer could
costs are similar for cases 1a and 2, but case 1b has factor 5 deviate considerably. The sensitivity analysis clearly shows
higher yearly costs, because of the investments in the heat that the distance has a large influence on the LCOE, especially
pump. This effect is reflected in the LCOE, which is 36.9V/ with direct heat delivery (case 1a, Fig. 8a) and delivery to a
MWhheat, while case 1a has an LCOE of 8.9 V/MWhheat. In case low-temperature DHN (case 2, Fig. 8c). Delivering waste heat
2, the LCOE is slightly lower than for 1a with 8.4 V/MWhheat, over long distances will thus not be feasible. The CO2 price is
because more heat can be utilised in this scenario due to the
larger heat buffer.
Table 10 e Costs with CO2 pricing - results for the cases.

Case 1a Case 1b Case 2
Table 9 e Costs without CO2-pricing - results for the
cases. Total yearly costs (kV/y) 19.7 106.5 19.7
CO2 emissions reduction 37.2 31.4 11.5
Case 1a Case 1b Case 2
cost savings (kV/year)
Total yearly costs (kV/y) 19.7 106.5 19.7 LCOEheat (heat production 7.9 26.0 3.5
LCOEheat (heat production 8.9 36.9 8.4 costs in euro/MWh)
costs in euro/MWh)
Saved yearly costs (kV/y) 84.2 109.4 38.3
Saved yearly costs (kV/y) 84.2 109.4 38.3 Margin including CO₂ emissions 101.8 34.3 30.0
Margin (kV/y) 64.5 2.9 18.6 reduction cost savings (kV/y)
Fig. 8 e Results of the sensitivity analysis with tornado diagrams of case 1a (a), case 1b (b) and case 2 (c).
the second most important factor in case 1a and 2 that in- transported, the lower the investment costs per unit of heat
fluences the LCOE, yet the adoption of a CO2 price would will be, as well as smaller heat losses [52,53].
positively influence the business case. When a heat pump is The sensitivity of the gas price is shown with respect to the
installed (case 1b, Fig. 8b), the electricity prices are the most margin in Fig. 9b. We have assumed a relatively low gas price
important factor in determining the LCOE. Long-term con- in the analysis based on historical data, but the war in Ukraine
tracts for electricity prices will give more certainty in the has caused a tremendous increase in gas prices. It is uncertain
business case. Another option would be to have dedicated how the gas price will develop in the coming years, yet it is
renewable production capacity for the heat pump electricity more likely to stay high. As expected, high gas prices will have
demand. In cases 1a and 2 (a&c), the electricity price is not a (highly) positive influence on the margin. In the analysis, the
such a major factor as electricity is only used for pumping the gas price and electricity price are changed by the same per-
heat around. Besides the distance, the percentage of over- centage, as the gas price will influence the electricity price.
lapping working hours shows to be sensitive as well in the Because of the electricity consumption of the heat pump in
case of delivery to a heat consumer (a&b). For case 1b (b), the case 1b, it has a lower margin than case 1a over the total
LCOE furthermore proves to be sensitive to the heat pump range.
temperature and to the CO2 price as well. Of slightly less
importance in all cases is the discount rate chosen.
The exact percentage of annual efficiency degradation, Discussion
costs for electronics and tie-in in the electrolyser for heat re-
covery have a high degree of uncertainty still. Yet, the sensi- General reflections and limitations of the study
tivity analysis shows that their impact on the LCOE remains
small. So, although we need more exact information about The results presented in this study are subjected to certain
these parameters, they are not the most essential while limitations. The lack of literature and concrete examples of
developing a business case for heat recovery from electrolysis. heat recovery from electrolysis, as mentioned in the intro-
Of these three parameters, the annual efficiency degradation duction as a reason to perform this study, is also its limitation.
has the largest impact on the LCOE. We have given insight into the feasibility of using electrolyser
As the distance shows to be an important factor in the waste heat, but we have had to make assumptions as there is
business case of electrolyser heat, Fig. 9a shows more pre- no available data. Therefore, we have included a sensitivity
cisely how the distance influences the margin between costs analysis (see also 4.1.3), which has shown that uncertain pa-
and benefits (saved costs) for the consumer. If heat is deliv- rameters such as the actual efficiency degradation and exact
ered directly to the consumer (1a) the maximum feasible prices for the tie-in and electronics have a limited influence on
distance is around 3 km. For heat delivery to the DHN (2), this the business case. Thus, the results obtained are reliable
distance is 1 km. For heat delivery with the heat pump (1b), the enough regardless of the uncertainties in these parameters.
margin is too small resulting in a negative business case above Another general comment is that every case will have its
200 m distance. These values are valid for an electrolyser size specific circumstances. The BOP will be different per supplier
of 2.5 MWel, with larger installations, the more heat is and will also depend on the engineering. Yet, the design we
Fig. 9 e Sensitivity of a) distance between electrolyser and heat consumer and b) gas price on the margin between costs and
saved costs at the heat consumer. The y-axis crossing shows the currently chosen value in the analysis. In b) the gas price is
shown on the x-axis, but in this analysis, the electricity price is changed simultaneously with the same percentage for a fair
comparison.
propose gives both possibilities to reuse heat within the (hydrogen and heat) are connected for the owner of the elec-
electrolyser as well as supply heat to an external consumer, trolyser, and thus the hydrogen production costs could play a
and could thus be helpful as a starting point for other studies. role in the business case of the waste heat. For example, in the
Regarding the economics, we should note that we have results (see 3.5), we have shown that high gas prices will in-
worked with a stable hydrogen production rate for over ten crease the margin between the costs of waste heat delivery
years. New electrolyser projects are not likely to start pro- and saved costs for the consumer. But, high gas and (there-
ducing at full capacity from year one. Over time the produc- fore) electricity prices will affect the business case for green
tion capacity will increase, building up to full capacity, thus in hydrogen and may lead to the hydrogen producer asking a
the first circa five years probably less heat will be delivered higher price for the waste heat to compensate for higher
(and less CO2 reduced). More precise calculations for specific electricity costs. This is even the case for green hydrogen
cases should be done to determine the exact effect on the produced with a solar or wind-based power purchase agree-
business case. For the cases presented in this paper, it means ment (PPA), as those PPA prices are affected by the wholesale
that the LCOE and margins presented should be seen as a first- market including fossil-based electricity sources [55].
order approximation. A highly feasible case (1a) will probably
stay feasible even with these start-up effects taken into ac- Ancillary services and hot standby mode
count. Furthermore, the decision for a heat consumer to
replace natural gas with waste heat does not have to be of a Electrolysers can fulfil ancillary services [56,57]. This would
purely economic nature, or more indirectly. A ‘green’ profile mean that the electrolyser would quickly (within seconds)
can enhance the reputation of a company, leading to a better reduce or increase its capacity [58]. We have not considered
position in the market and eventually to higher financial the impact of this mode of operation in this study. Yet,
performance [54]. although the ancillary services can have a positive impact on
Furthermore, in this study, we only considered the costs to the business case of the electrolyser [59], the total amount of
extract and deliver waste heat, independent from the hydrogen (and thus heat) produced will not vary significantly
hydrogen production costs. Eventually, both ‘products’ when production targets have to be met. We thus expect that
the influence on the amount of heat produced will not deviate

Table 11 e Heat source cost (LCOE) comparison.
more than a few per cent.
When ancillary services are part of the business case for an Heat source Costs (LCOE)
in V/MWhheat
electrolyser, it either has to operate at the time when it is
bidding in for the balancing markets, or it has to stay in hot Waste heat electrolyser to DHN e 8.4e8.9
standby mode so it can react quickly [58]. In the hot standby this publication without heat
pump (54 C)
mode, the electrolyser will consume electricity, about 3% of its
Waste heat electrolyser to DHN e 36.9
nominal power [58,60]. Part of the electricity consumption for
this publication with heat pump
the hot standby mode is used to keep the electrolyser warm (100 C)
and could maybe be supplied by its own waste heat. With the Heat from surface water with a 55.4e73.4
assumption of 50% of the hot standby energy use being ful- heat pump (45 C)
filled by waste heat, we calculated that about 28 MWh/year of Industrial waste heat sources 36e46
heat could be used within the electrolyser itself (for case 1a). (<35 C, with heat pump to
75 C) [62]
This is a small potential but could be interesting in combina-
Industrial waste heat sources 6e10
tion with other waste heat uses, so the costs for a heat buffer (75e100 C) [62]
could be shared. Geothermal energy at low 54e65
temperature (50 C, depth 500
Efficiency raise of electrolyser stack e1000 m, including heat
pump) [63]
We have seen an efficiency raise of the total PEM electrolyser
stack from 76% to 90/91% (HHV based) when the electrolyser heat sources below 40 C, a heat pump is necessary to increase
heat is taken into account, thus a 14e15% efficiency increase the temperature, which is reflected in the LCOE. The LCOE of
in total. Bilbao mentions a 13% potential efficiency increase electrolyser heat without a heat pump including a short
for an alkaline electrolyser if heat would be recovered [21]. In transport distance (8.4e8.9 V/MWhheat) is well below this
Buttler & Spliethoff two projects are mentioned that increase range (36e73.4 V/MWhheat). Even with a heat pump to increase
their efficiency including heat utilisation to 95% and 86%. the temperature, the costs for electrolyser heat are at the
Frank et al. [23] show a range of options for a 1 MW Power-to- lower end of the range with 36.9 V/MWhheat, while the tem-
Gas plant with an alkaline electrolyser including different perature is higher (100 C). Electrolyser heat could compete as
parts of the electrolyser design to recover heat from. With the well with higher temperature waste heat sources, but only at a
stack, electrolyte circuit, product gas treatment and water lower temperature level (at 54 C), so without a heat pump.
supply taken into account, they calculated an efficiency gain Overall, these data show that the waste heat from the elec-
of external heat use from circa 78% to 97% with a 60 C oper- trolyser may compete with other industrial waste heat sour-
ating temperature. For the overall electrolysis system, the ces depending on the necessary temperature level, and is
efficiency gain would be 17.6% (from 72.8 to 90.4%). Jonsson & likely to be cheaper than lower-temperature heat sources.
Miljanovic found an overall system efficiency for a PEM elec- Other sources that specifically discuss the business case of
trolyser including heat recovery of 94.7%, an 18% increase heat from electrolysers are scarce. Jonsson & Miljanovic found
from their reference scenario if all waste heat can be utilised costs of 20 V/MWhth for waste heat utilisation from a PEM
(at 80 C working temperature) [61]. So although most avail- electrolyser of 100 MW, with a heat pump raising the tem-
able data are for alkaline electrolysis, our findings for the ef- perature to 120 C [61]. Without the heat pump (so only a heat
ficiency raise are comparable with the literature. Moreover, exchanger), they calculated 1.65 V/MWhth, yet costs for a tie-
we have not yet considered the possible reduction in energy in and controls have not been taken into account here so
use of the dry coolers of the electrolyser system, which could these values seem to be an underestimation. Our results show
lead to lower energy use of the electrolysis process. The heat higher costs because more cost components have been taken
utilisation from the electrolyser can thus clearly increase the into account, but Jonsson & Miljanovic [61] show that with
overall system efficiency of the electrolysis process. When larger electrolyser installations, the costs will probably
applied, the electrolyser will be a hydrogen as well as heat decrease.
producer and will help to make even better use of available
renewable energy. Theoretical potential of electrolyser heat
Comparison to other (waste) heat costs In our case study, we have considered a 2.5 MW electrolyser
that could part of a local hydrogen cluster. Local green
We have calculated the LCOE of waste heat from the electro- hydrogen production is expected to grow exponentially in the
lyser for different temperature levels and applications. We coming years and decades, within the current plans of for
have concluded that there is a high possibility for a positive example the EU and its RePowerEU programme [5]. The waste
business case for direct heat delivery to a heat consumer and heat from the electrolyser could thus potentially become a
heat delivery to a low-temperature DHN (40e60 C). Moreover, waste heat source for district heating. Yet, there are many
it is relevant to compare our results with other sources of other possible waste heat sources available, so what would be
(waste) heat, especially for district heating networks. Table 11 the share of heat from hydrogen production compared to the
shows a comparison of heat production costs per MWh, total available waste heat potential? We will focus on Europe
without the distribution costs. For low-temperature waste- to answer this question. Fleiter et al. [64] calculated that there
is a potential of 425 PJ (118 TWh th) of heat at 95 C available as well, including different modes of operation of the elec-
within Europe, 503 PJ (140 TWh th) at 55 C and 960 PJ (267 TWh trolyser as well as ancillary services. Furthermore, there is the

th) with a temperature of 25 C. Of this 425 PJ at 95 C, 151 PJ possibility to use part of the waste heat for the hot standby
could be utilised within a 10 km range from the heat source in mode of the electrolyser, combined with another heat
existing district heating grids and almost all of the heat (415 PJ consumer.
e 115 TWh th) would be useable considering the expected in-
crease of district heating grids.
How does the amount of heat from electrolyser in- Conclusions
stallations relate to this? The EU target is 10 million tonnes of
domestic hydrogen production in 2030 [5]. According to van We have shown that it is possible to make a redundant system
Wijk et al. [65], about 5% of this hydrogen production capacity design to utilise heat from an electrolyser without compro-
will be decentral capacity, and 15% will be produced directly at mising on the hydrogen production process. The utilisation
the industry/point of use itself (captive market). The other 80% and valorisation of heat from electrolysers could lead to.
of green hydrogen production will be large centralized pro-
duction, at places with good solar and wind resources, prob- Higher system efficiency - An increase in combined electro-
ably further away from (heat) consumers. For the captive lyser system (stack þ waste heat) efficiency from 76% to
market and decentral capacity, it is most likely that there is a 90e91%, based on HHV. Furthermore, 14e15% of the elec-
useful purpose for the waste heat. Thus, if we assume that tricity input to the electrolyser stack can be utilised as heat
about 20% of the expected hydrogen production capacity by a heat consumer, depending on the use case. For the
could lead to useful heat production, this corresponds to 2 district heating system (case 2), we have shown that the
million tonnes of hydrogen production. With 80% efficiency of electrolyser (waste) heat can fulfil around 16% of the total
the electrolysis (HHV based), this amounts to 98.5 TWh of heat demand of the neighbourhood.
electricity input. We have calculated that 14% of the electricity CO2 savings - Direct heat use (case 1a) leads to 0.28 tonne
input will be used as heat (see section 3.2), so around 13.8 CO2/MWhheat, used. With a heat pump (case 1b) to increase
TWhth of electrolyser heat could be added to the total waste the heat temperature, the savings would be 0.18 tonne CO2/
heat potential within the EU. This is about 10% of the waste MWhheat, used if the heat pump works on an average elec-
heat potential at 55 C, a modest, but not insignificant amount tricity grid mix. Delivery of heat to a low-temperature DHN
of heat. Bo € hm et al. calculated a 56e84 TWhth (theoretical) leads to 0.08 tonne CO2/MWhheat, used, compared to heat
potential of thermal energy from hydrogen supply based on produced directly by a heat pump working on an average
the European hydrogen strategy [26], but haven't corrected for electricity grid mix.
hydrogen production in proximity of heat demand. Our cal- Economic feasibility - When a heat pump is necessary to in-
culations are thus probably closer to the technical potential of crease the waste heat temperature, it will be harder to
hydrogen waste heat. realise a business case (LCOE 36.9 V/MWhheat, used) than
without a heat pump (8.9 V/MWhheat, used). If CO2-pricing of
Recommendations for future research 60 V/tonne would be taken into account, the business case
is likely to be positive in all cases (LCOE of 7.9 - 26
We have obtained insights on how heat from electrolysers V/MWhheat, used). The sensitivity analysis has shown that
could (technically) be recovered as well as an energetic, the transport distance of heat is an important factor in
environmental (CO2-reduction) and economic potential. Yet, determining the feasibility of electrolyser heat utilisation.
the adoption of heat utilisation from electrolysers is still small The feasible distance varies per use case between 1 and
so there is a lack of real data from actual projects. Validation 3 km. Compared to other waste heat sources, the LCOE of
of results is now merely done with other theoretical studies, the electrolyser heat (8.4e36.9 V/MWhheat, used) falls within
so for future research, validation with practical data would be or below the range of lower-temperature heat sources.
very valuable. These data could also give more insight into a Without a heat pump, the LCOE (8.4e8.9 V/MWhheat, used)
stepwise increasing hydrogen (and thus heat) production also falls within the range of other high-temperature in-
pattern during the first period of operation. Furthermore, dustrial heat sources.
other use cases can be evaluated. Some interesting other use
cases could be a wastewater treatment plant with local Overall, we show that electrolyser heat can both from an
hydrogen production (where oxygen could be utilised in environmental and economic point of view be a valuable
aeration) or a hydrogen fuelling station with local hydrogen addition to a local integrated energy system, and further
production and use of the waste heat for car washing. Another enhance local system optimisation by integrating both
possibility is to use the waste heat for thermal desalination, hydrogen and heat production.
thereby producing deionized water at lower costs than with
reversed osmosis technology [66].
If a heat pump is necessary to increase the waste heat Data
temperature, more research could be done on how local
renewable capacity would influence the business case in Raw data of the modelling results from the cases described in
terms of electricity prices. Lastly, the results could be further the publication can be accessed via DOI: https://doi.org/10.
refined by taking into account the business case of hydrogen 4121/21666152.
[7] van der Zwaan B, Lamboo S, Dalla Longa F. Timmermans'

Funding dream: an electricity and hydrogen partnership between
Europe and North Africa. August Energy Pol 2021;159:112613.
The H-Flex project is funded by the Dutch National Chamber https://doi.org/10.1016/j.enpol.2021.112613 [Online].
of Commerce (RVO) via the DEI programme (Demonstration [8] van Wijk A, Wouters F. Hydrogen - the bridge between Africa
and Europe. In: Weijnen MCP, Lukszo Z, Farahani S, editors.
Energy Innovation) with project number DEI4891013.
Shaping an inclusive energy transition. Springer Cham; 2021.
The electrolyser in the H-Flex project is operated by Hys-
p. 91e120 [Online]. Available, https://link.springer.com/
olar (https://www.hysolar.nl/en) and co-funded by the EU LIFE chapter/10.1007/978-3-030-74586-8_5.
programme via the LIFE NEW HYTS project (LIFE20 CCM/NL/ [9] Gas for Climate and Guidehouse, “Extending the European
001664). Hydrogen Backbone; A European Hydrogen Infrastructure
vision covering 21 countries,” 2021 [Online]. Available:
https://www.ehb.eu/files/downloads/European-Hydrogen-
Backbone-April-2021-V3.pdf
Declaration of competing interest
[10] van Wijk A, Westphal K, Braun JF. How to deliver on the EU
hydrogen accelerator [Online]. Available, https://
The authors declare that they have no known competing hydrogeneurope.eu/wp-content/uploads/2022/05/How-to-
financial interests or personal relationships that could have deliver-on-the-EU-Hydrogen-Accelerator_Final.pdf.
appeared to influence the work reported in this paper. [11] Siemens Energy. Silyzer PEM electrolysers, an overview. 2020
[Online]. Available, https://4echile-datastore.s3.eu-central-1.
amazonaws.com/wp-content/uploads/2020/10/10132733/
20200930-SE-NEB-PEM-Electrolyzer-and-Applications_EW.
Acknowledgements pdf.
[12] Cummins. HyLYZER® WATER ELECTROLYZERS. Cummings
The authors would like to thank Hysolar for their collabora- Product Specification Sheets 2021:30e1 [Online]. Available,
https://www.cummins.com/sites/default/files/2021-08/
tion in the project and for providing information about the
cummins-hylyzer-1000-specsheet.pdf.
electrolyser. The authors would like to thank Hans Huiting
[13] Buttler A, Spliethoff H. Current status of water electrolysis
from Hans Huiting Onderzoek & Advies for his help on the for energy storage, grid balancing and sector coupling via
initial design of the balance of plant of the electrolyser with a power-to-gas and power-to-liquids: a review. September
tie-in. The authors would also like to thank Danie €l Bakker Renew Sustain Energy Rev 2017;82:2440e54. https://doi.org/
from KWR Water for his contributions to the economic model 10.1016/j.rser.2017.09.003. 2018.
for the waste heat calculations. [14] Power-to-Gas via biological catalysis - final report; 2014.
[Online]. Available https://energiforskning.dk/en/node/
15072.
[15] IKZ, “Der „Stromlückenfüller“,” 2017. [Online]. Available:
Appendix A. Supplementary data https://www.ikz.de/detail/news/detail/der-
stromlueckenfueller/. [Accessed: Apr. 12, 2023]
Supplementary data to this article can be found online at [16] Gmbh GHE. Green hydrogen esslingen [Online]. Available,
https://doi.org/10.1016/j.ijhydene.2023.03.374. https://green-hydrogen-esslingen.de/. [Accessed 1 March
2023].
[17] RWE starts up P2G plant with ITM electrolyser. Fuel Cell Bull
references 2015;2015(8):1. https://doi.org/10.1016/S1464-2859(15)30205-4.
[18] Shell, Mitsubishi heavy industries, Vattenfall and Wa € rme
Hamburg sign letter of intent for 100MW hydrogen project in
Hamburg. 2021 [Online]. Available, https://www.hghh.eu/
[1] IRENA. Global renewables outlook: energy transformation news/shell-mitsubishi-heavy-industries-vattenfall-and-
2050. 2020 [Online]. Available, https://www.irena.org/ warme-hamburg-sign-letter-of-intent-for-100mw-
publications/2020/Apr/Global-Renewables-Outlook-2020. hydrogen-project-in-hamburg.
[2] DNV. HYDROGEN FORECAST TO 2050 energy transition [19] Yuan M, Thellufsen JZ, Sorknæs P, Lund H, Liang Y. District
outlook 2022. 2022 [Online]. Available, https://www.dnv.com/ heating in 100% renewable energy systems: combining
focus-areas/hydrogen/forecast-to-2050.html. industrial excess heat and heat pumps. Energy Convers
[3] Hydrogen Council. How hydrogen empowers the energy Manag 2021;244(18):114527. https://doi.org/10.1016/
transition. 2017 [Online]. Available, https://hydrogencouncil. j.enconman.2021.114527 [Online].
com/wp-content/uploads/2017/06/Hydrogen-Council-Vision- [20] van ’t Noordende H, Ripson P. A one-GigaWatt green-
Document.pdf. hydrogen plant - advanced design and total installed-capital
[4] Fuel Cells and Hydrogen Joint Undertaking (FCH). Hydrogen costs. 2020 [Online]. Available, https://ispt.eu/media/ISPT-
roadmap Europe. 2019 [Online]. Available, https://www.fch. public-report-gigawatt-green-hydrogen-plant.pdf.
europa.eu/sites/default/files/Hydrogen Roadmap Europe_ [21] Contreras Bilbao D. Valorization of the waste heat given off
Report.pdf. in a system alkaline electrolyzer-photovoltaic array to
[5] EU Commission. REPowerEU: joint European Action for more improve hydrogen production performance: case study
affordable, secure and sustainable energy [Online]. Available, Antofagasta, Chile. Int J Hydrogen Energy
https://eur-lex.europa.eu/legal-content/EN/TXT/?uri¼COM: 2021;46(61):31108e21. https://doi.org/10.1016/
2022:108:FIN; 2022. j.ijhydene.2021.07.016 [Online].
[6] European Commission. A hydrogen strategy for a climate- [22] Hu Q, Lin J, Zeng Q, Fu C, Li J. Optimal control of a hydrogen
neutral Europe, Brussels, pp. 1-23, 2020 [Online]. Available, microgrid based on an experiment validated P2HH model.
https://eur-lex.europa.eu/legal-content/EN/TXT/?uri¼CELEX: IET Renew Power Gener 2020;14(3):364e71. https://doi.org/
52020DC0301. 10.1049/iet-rpg.2019.0544.
[23] Frank E, Gorre J, Ruoss F, Friedl MJ. Calculation and analysis Energies May 2021;14(9):2616. https://doi.org/10.3390/
of efficiencies and annual performances of Power-to-Gas en14092616 [Online]. Available, https://www.mdpi.com/
systems. Appl Energy 2018;218(February):217e31. https:// 1996-1073/14/9/2616.
doi.org/10.1016/j.apenergy.2018.02.105 [Online]. [38] van der Roest E, Beernink S, Hartog N, van der Hoek JP,
[24] Huang C, Zong Y, You S, Træholt C. Economic model Bloemendal M. Towards sustainable heat supply with
predictive control for multi-energy system considering decentralized multi-energy systems by integration of
hydrogen-thermal-electric dynamics and waste heat subsurface seasonal heat storage. Energies Nov.
recovery of MW-level alkaline electrolyzer. Energy Convers 2021;14(23):7958. https://doi.org/10.3390/en14237958 [Online].
Manag 2022;265(May):115697. https://doi.org/10.1016/ Available, https://www.mdpi.com/1996-1073/14/23/7958.
j.enconman.2022.115697 [Online]. [39] Schilling J, van de Poll F, van Berkel P. CO2-effecten van de
[25] Huckebrink D, Bertsch V. Decarbonising the residential Transitievisie Warmte: Gemeente amsterdam. Delft, pp. 1-
heating sector: a techno-economic assessment of selected 23, 2020 [Online]. Available, https://ce.nl/wp-content/
technologies. Energy 2022;257:124605. https://doi.org/ uploads/2021/03/CE_Delft_200154_-CO2-effecten_
10.1016/j.energy.2022.124605 [Online]. Transitievisie_Warmte_Amsterdam_DEF.pdf.
[26] Bo€ hm H, Moser S, Puschnigg S, Zauner A. Power-to-hydrogen [40] IF Technology. Thermische Energie uit Oppervlaktewater -
& district heating: technology-based and infrastructure- Business case ‘Genderdal’ Eindhoven. 2018 [Online].
oriented analysis of (future) sector coupling potentials. Int J Available, https://edepot.wur.nl/460776.
Hydrogen Energy 2021;46(63):31938e51. https://doi.org/ [41] Stadler M, Aki H, Firestone R, Lai J, Siddiqui A. Distributed
10.1016/j.ijhydene.2021.06.233. energy resources on-site optimization for commercial
[27] van der Roest E, Vries D, van Wijk AJM. H-Flex: green buildings with electric and thermal storage technologies.
hydrogen in a flexible electricity system. KWR Water 2020 2008 ACEEE summer study energy effic. Build. Scaling up
[Online]. Available, https://www.kwrwater.nl/en/projecten/ build. Tomorrow's solut., 8; May 2008. p. 1e12. 2008 [Online].
h-flex-green-hydrogen-in-a-flexible-electricity-system/. Available, https://escholarship.org/uc/item/8nv7m8sx.pdf.
[Accessed 16 March 2022]. [42] EnergiNet and T. D. E. Agency. Technology data energy
[28] Mancera JJC, Manzano FS, Andújar JM, Vivas FJ, Caldero n AJ. storage. 2020 [Online]. Available, https://ens.dk/en/our-
An optimized balance of plant for a medium-size PEM services/projections-and-models/technology-data/
electrolyzer. Design, control and physical implementation. technology-data-energy-storage.
Electron 2020;9(5). https://doi.org/10.3390/ [43] van der Roest E, Snip L, Fens T, van Wijk A. Introducing
electronics9050871. Power-to-H3: combining renewable electricity with heat,
[29] Tiktak WJ. Heat management of PEM electrolysis. 2019 water and hydrogen production and storage in a
[Online]. Available, http://resolver.tudelft.nl/uuid:c046820a- neighbourhood. Appl Energy Jan. 2020;257. https://doi.org/
72bc-4f05-b72d-e60a3ecb8c89. 10.1016/j.apenergy.2019.114024 [Online].
[30] Kopp M, Coleman D, Stiller C, Scheffer K, Aichinger J, [44] SKAO, Stimular, Connekt, Milieu Centraal, and Ministerie van
Scheppat B. Energiepark Mainz: technical and economic Infrastructuur en Milieu. CO2 emissiefactoren. 2022 [Online].
analysis of the worldwide largest Power-to-Gas plant with Available, https://co2emissiefactoren.nl/. [Accessed 16
PEM electrolysis. Int J Hydrogen Energy 2017;42(19):13311e20. November 2022].
https://doi.org/10.1016/j.ijhydene.2016.12.145 [Online]. [45] CBS. Aardgas en elektriciteit, gemiddelde prijzen van
[31] Thyssenkrupp. Hydrogen from large-scale electrolysis. 2018 eindverbruikers. 2021 [Online]. Available, https://opendata.
[Online]. Available, https://www.thyssenkrupp-uhde- cbs.nl/#/CBS/nl/dataset/81309NED/table?dl¼6ABB2.
chlorine-engineers.com/en/products/water-electrolysis- [Accessed 29 June 2022].
hydrogen-production/power-to-gas/. [46] OECD. “Effective Carbon Rates 2021.” pp. 1e12, 2021 [Online].
[32] Davies J, Dolci F, Weidner E. Historical Analysis of FCH 2 JU Available: https://www.oecd.org/tax/tax-policy/effective-
Electrolyser Projects: evaluation of contributions towards carbon-rates-2021-brochure.pdf
advancing the state of the art. EUR 30648. Luxembourg: [47] Ember. Carbon Price Tracker - the latest data on EU and UK
Publications Office of the European Union, 2021 [Online]. ETS carbon prices [Graph]. 2023 [Online]. Available, https://
Available, https://publications.jrc.ec.europa.eu/repository/ ember-climate.org/data/data-tools/carbon-price-viewer/.
bitstream/JRC121704/electrolyser_hist._analysis_public_ [Accessed 1 March 2023].
20200323final_identifiers_1.pdf. [48] Polytherm. Complete price list March 2019. 2019 [Online].
[33] Nederlands Normalisatie-Instituut. NEN-EN 13941-1-District Available, http://www.polytherm.ie/v4/0940aa0c-5421-4a9b-
heating pipes - design and installation of thermal insulated 840d-c9a2ae5d95bb/uploads/PolythermCompletePricelist.
bonded single and twin pipe systems for directly buried hot pdf.
water networks. 2016. [Online]. Available https://www.nen. [49] EnergiNet and The Danish Energy Agency. Technology data -
nl/nen-en-13941-1-2019-en-258340. generation of electricity and district heating. 2020 [Online].
[34] Arpagaus C, Bless F, Uhlmann M, Schiffmann J, Bertsch SS. Available, https://ens.dk/en/our-services/projections-and-
High temperature heat pumps: market overview, state of the models/technology-data/technology-data-generation-
art, research status, refrigerants, and application potentials. electricity-and.
Energy 2018;152:985e1010. https://doi.org/10.1016/ [50] Task 1: technologies e State of the art and ongoing
j.energy.2018.03.166. developments for systems and components. Annex 58 about
[35] European Heat Pump Association. Large scale heat pumps in HTHP 2022 [Online]. Available, https://
Europe, 2; 2020. [Online]. Available https://www.ehpa.org/ heatpumpingtechnologies.org/annex58/task1/. [Accessed 24
wp-content/uploads/2022/11/Large-heat-pumps-in-Europe- February 2023].
and-industrial-uses_2020.pdf [51] Croezen H, Doe M, Jorna N. Broeikasgasemissies
[36] Lund H, et al. Perspectives on fourth and fifth generation aardgasketens. 2022 [Online]. Available, https://www.
district heating. Energy 2021;227:120520. https://doi.org/ co2emissiefactoren.nl/wp-content/uploads/2022/01/35-
10.1016/j.energy.2021.120520 [Online]. RHDHV-Broeikasgasemissies-aardgasketens-2022.pdf.
[37] van der Roest E, Fens T, Bloemendal M, Beernink S, van der [52] Colmenar-Santos A, Rosales-Asensio E, Borge-Diez D,
Hoek JP, van Wijk AJM. The impact of system integration on Blanes-Peiro J-J. District heating and cogeneration in the EU-
system costs of a neighborhood energy and water system. 28: current situation, potential and proposed energy strategy
for its generalisation. Renew Sustain Energy Rev [59] HyBalance. Final report with strategic learnings from
2016;62:621e39. https://doi.org/10.1016/j.rser.2016.05.004 performance reporting, business case analysis [Online].
[Online]. Available, https://cordis.europa.eu/project/id/671384/results;
[53] Kavvadias KC, Quoilin S. Exploiting waste heat potential by 2021.
long distance heat transmission: design considerations and [60] Matute G, Yusta JM, Beyza J, Correas LC. Multi-state techno-
techno-economic assessment. February Appl Energy economic model for optimal dispatch of grid connected
2018;216:452e65. https://doi.org/10.1016/ hydrogen electrolysis systems operating under dynamic
j.apenergy.2018.02.080 [Online]. conditions. Int J Hydrogen Energy 2021;46(2):1449e60. https://
[54] Baah C, et al. Examining the correlations between doi.org/10.1016/j.ijhydene.2020.10.019 [Online].
stakeholder pressures, green production practices, firm [61] Jonsson F, Miljanovic A. Utilization of waste heat from
reputation, environmental and financial performance: hydrogen production. Ma € lardalen University; 2022 [Online].
evidence from manufacturing SMEs. Sustain Prod Consum Available, https://www.diva-portal.org/smash/get/diva2:
2021;27:100e14. https://doi.org/10.1016/j.spc.2020.10.015 1670187/FULLTEXT01.pdf.
[Online]. [62] Muller M, Lensink S. Conceptadvies SDEþþ CO2-reducerende
[55] Stones J, Hamilton J. ICIS WHITEPAPER: renewable PPAs and opties. 2019 [Online]. Available, https://www.pbl.nl/sites/
a review of the commodity price spike on renewable default/files/downloads/pbl-2019-conceptadvies-SDE-plus-
hydrogen production costs. ICIS 2022:1e6 [Online]. Available, plus-industriele-restwarmte_3745.pdf.
https://www.icis.com/explore/resources/news/2022/10/11/ [63] Schepers B, Scholten T, Willemsen G, Koenders M, de
10813598/icis-whitepaper-renewable-ppas-and-a-review-of- Zwart B. Weg van gas - Kansen voor de nieuwe concepten
the-commodity-price-spike-on-renewable-hydrogen- LageTemperatuurAardwarmte en Mijnwater. 2018 [Online].
production-costs/. Available, https://ce.nl/wp-content/uploads/2021/03/CE_
[56] International Renewable Energy Agency (IRENA). Demand- Delft_3K61_Weg_van_gas_DEF.pdf.
SIDE flexibility for power sector. Abu Dhabi. 2019 [Online]. [64] Fleiter T, et al. Excess heat potentials of industrial sites in
Available, https://www.irena.org/-/media/Files/IRENA/ Europe, Documentation of excess heat potentials of industrial
Agency/Publication/2019/Dec/IRENA_Demand-side_ sites including open data file with selected potentials. 2020.
flexibility_2019.pdf. p. 1e76 [Online]. Available, https://www.seenergies.eu/wp-
[57] Sua rez VG, Rueda Torres JL, Tuinema Arcadio BW, Guerra P, content/uploads/sites/25/2020/04/sEEnergies-WP5_D5.1-
Van Der Meijden M. Integration of power-to-gas conversion Excess_heat_potentials_of_industrial_sites_in_Europe.pdf.
into Dutch electricity ancillary services markets. In: 12th [65] Van Wijk A., Chatzimarkakis J. Hydrogen Europe sets out
conference on energy economics and technology; 2018. 240 GW green hydrogen initiative, Fuel Cells Bull., no. 5, pp.
p. 1e7 [Online]. Available, http://resolver.tudelft.nl/uuid: 10e11, May 2020, doi: 10.1016/S1464-2859(20)30204-2.
d63402bb-2e26-477c-b3d7-e3c6583242ff. [Online]. Available: http://www.magonlinelibrary.com/doi/
[58] Matute G, Yusta JM, Correas LC. Techno-economic modelling 10.1016/S1464-2859%2820%2930204-2
of water electrolysers in the range of several MW to provide [66] Ginsberg M, Zhang Z, Atia AA, Venkatraman M, Esposito DV,
grid services while generating hydrogen for different Fthenakis VM. Integrating solar energy, desalination, and
applications: a case study in Spain applied to mobility with electrolysis. Sol. RRL 2022;6(5):1e13. https://doi.org/10.1002/
FCEVs. Int J Hydrogen Energy 2019;44(33):17431e42. https:// solr.202100732.
doi.org/10.1016/j.ijhydene.2019.05.092 [Online].

1 s2.0 S0360319923015410 Main

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

1 s2.0 S0360319923015410 Main

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

1 s2.0 S0360319923015410 Main

Uploaded by

Copyright:

Available Formats

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 2 7 8 7 2 e2 7 8 9 1

Available online at www.sciencedirect.com

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

Utilisation of waste heat from PEM electrolysers e

highlights graphical abstract

57°C 90% 0.28 8.9

62°C 57°C 100°C

heating grid. Electrolyser

The feasibility of heat utilisation is

very sensitive to the transport

article info abstract

The European Commission has raised the ambition of

Heat sources in an electrolyser Variance in heat production

Efficiency & energy balance PNt PNt

sumption of the electrolyser in kWh over the period i to Nt in

Where LCi represents the annual levelized costs for a certain

Fig. 4 e Redundant designs for heat recovery from a PEM electrolyser.

Table 8 shows the simulation results for all cases,

savings utilised effi-ciency demand lyser heat savings

year) pump (MWh)

and the overall efficiency are shown in Fig. 7. Eventually,

electricity input in the electrolyser is used as heat by a heat

crease the temperature, thereby increasing the DT and thus

the amount of heat available at the consumer. The efficiency

In case 2, the electrolyser heat is reused in a low-

electrolyser is used in a low-temperature DHN. The amount of

heat provided increased over the years and accounts on

percentage of fulfilled heat demand by the electrolyser is

caused by a difference in heat demand over the years due to

used as heat in the DHN. Interesting to note is that the elec-

Table 10 e Costs with CO2 pricing - results for the cases.

the influence on the amount of heat produced will not deviate

[7] van der Zwaan B, Lamboo S, Dalla Longa F. Timmermans'

You might also like