DYNAMIC MODELING AND SIMULATION OF ALKALINE TYPE ELECTROLYZERS

Md. Maruf-ul-Karim1, M. T. Iqbal

Faculty of Engineering and Applied Science

Memorial University of Newfoundland, St. John’s, NF, Canada, A1B3X5
1
Email: [email protected]

ABSTRACT model of an alkaline electrolyzer in this paper can be used to

study both steady state and transient operations. The HySTATTM-
Electrolyzers are essential part of power systems where the energy A 1000D-30-25 outdoor version electrolyzer is considered in this
storage is in the form of hydrogen. Electrolyzers are used in research. It will be installed with all its utility equipments at
remote hybrid power systems. In this research a dynamic model of Ramea, Newfoundland in 2009. The cell voltage, Faraday
an alkaline type electrolyzer has been developed. The MATLAB/ efficiency, energy efficiency, hydrogen production, heat loss,
Simulink software has been used to form the generalized dynamic auxiliary cooling demand and some other related parameters are
model, which can be integrated with different kinds of renewable studied and presented below graphically along with their time
energy systems. This paper describes the operation of electrolyzer dependency.
cell by the electrochemical and thermal models and presents the
simulation results graphically. This model is also suitable for 2. CHEMICAL REPRESENTATION
steady state and transient analysis of a remote power system.
A DC current is required to split the water molecule into
Index Terms – Alkaline type electrolyzer, Hydrogen storage, Fuel hydrogen and oxygen. This current should flow between two
cell systems, Dynamic modeling. electrodes (Anode and Cathode) separated by the electrolyte.
- +
1. INTRODUCTION
2e-
Environmentally clean energy production is an important issue 2e-
nowadays. There are various ways to produce renewable energy
and store that in a remote location. The solar and the wind are the H2 ½ O2
most common forms of energy that can be converted to electricity
to meet a load demand. A number of research projects have
2OH- 2OH-
already been carried out [3-7] and some are being continued. Electrolyte
(30% wt. KOH)
Those renewable energy sources are available in a certain amount
H2O H2O
at all territories. But due to their sporadic behavior, energy storage
is necessary.

Battery bank can be a solution of energy storage, but it has low Cathode Diaphragm Anode
energy density, self-discharge and leakage characteristics. It is also (Ni, C-Pt) (NiO) (Ni, Co, Fe)
not suitable for long-term energy storage [4]. As this paper focuses
on a stand-alone power system, which is located at Ramea, the
main focus is long-term energy storage. Hydrogen has the property Figure 1: Internal structure of an alkaline electrolyzer.
of high mass energy density which can be stored for long period
without energy loss. This makes hydrogen suitable for energy Here the electrolyte is 30% wt. KOH. Internal structure is shown
storage. Moreover, hydrogen is best suited where the conventional in figure 1. The reactions at anode, cathode and the total reaction
fossil fuel is expensive. for water splitting are:

The Proton Exchange Membrane (PEM) and Solid Oxide Anode reaction:
Electrolyzers (SOE) are promising configurations. But alkaline 1
electrolyzers are commercially available and it has been selected 2OH − (aq ) → O2 ( g ) + H 2 O(l ) + 2e − (1)
for Ramea wind-diesel-hydrogen project. The proposed simulink 2

978-1-4244-3508-1/09/$25.00 ©2009 IEEE 711

Authorized licensed use limited to: PAKISTAN INST OF ENGINEERING AND APPLIED SCIENCES. Downloaded on December 11,2022 at 19:08:39 UTC from IEEE Xplore. Restrictions apply.
 Cathode reaction: where I is the current required for the electrolysis, A is the area of
− − electrodes, r is the parameter related to ohmic resistance of the
2 H 2 O(l ) + 2e → H 2 ( g ) + 2OH (aq ) (2)
electrolyte, s and t are the over voltage coefficients on electrodes
and Vrev is the reversible voltage.
Total:
1 The reversible voltage can be expressed as,
H 2 O(l ) → H 2 ( g ) + O2 ( g ) (3)
ΔG
2 Vrev = (5)
ΔH=286 kJ/mole. zF
Reversible voltage is the minimum voltage required for starting the where ΔG is Gibb’s energy, z is the number of electrons
water electrolysis. In this study the value of reversible voltage is transferred in each reaction (here it is 2) and F is the Faraday
1.228V per cell. As water is not a good electric conductor, KOH is constant. In standard condition (25º C and 1 bar), the change of
added to increase the conductivity level of the electrolyte. Gibb’s energy of water splitting is ΔGº = 237 kJ/mol.

Anode is made of nickel (Ni), cobalt (Co) and iron (Fe) while, In addition to that, ΔH is the change of enthalpy between
cathode is made of nickel and platinum activated carbon catalyst reactants and products in a reaction and is related to thermo
(C-Pt) are used instead of normal electrodes to ensure that those neutral cell voltage as
can prevent corrosion in alkaline solution and have good electric
ΔH
conductivity. On the other hand the diaphragm should have low Vtn = (6)
electric resistance, that is why nickel oxide (NiO) is used as zF
diaphragm element. Figure 1 also shows the chemical reactions
taking place in the electrolyte. The energy required for water splitting in standard condition, ΔHº
= 286 kJ/mol.
3. ELECTROCHEMICAL MODEL
In order to relate the ohmic resistance parameter, r and over
The kinetic properties around the electrodes in an electrolyzer cell voltage coefficient, t the equation (4) can be modified in more
can be modeled based on empirical data of different proposed I-V detailed way as
curves. The basic relationship between current and cell voltage in a
given temperature is I
V = Vrev + (r1 + r 2 × T )
I I A
(4) (7)
V = Vrev + r + s log( t + 1) I t 2 t3
A A + s log[ (t1 + + 2 ) + 1]
A T T

Figure 2: Electrochemical simulation model.

712

Authorized licensed use limited to: PAKISTAN INST OF ENGINEERING AND APPLIED SCIENCES. Downloaded on December 11,2022 at 19:08:39 UTC from IEEE Xplore. Restrictions apply.
 Faraday efficiency is the proportion of the actual and theoretical 4. THERMAL MODEL
hydrogen production in the electrolyzer [1]. It will be shown later
that one of the parameters that control the hydrogen production is Like the electrochemical model, the cell temperature can be
current, so another name of Faraday efficiency is current calculated using either simple or complex thermal model.
efficiency. It can be noted that parasitic current increases with the Considering the electrolyzer as a lumped thermal capacitance
decrement of current density and the increment of temperature model the overall energy balance will be
resulting the reduction of Faraday efficiency. This phenomenon dT
can be expressed as Ct ′ − Qloss
= Q gen ′ − Qcool
′ (11)
dt
(I ) 2

ηF = A f2 (8) ′ = nc I (V − Vtn ) = ncVI (1 − η e )

Q gen (12)
f1 + ( I ) 2 1
A ′ =
Qloss (T − Ta ) (13)
where f1 and f2 are the parameters related to Faraday efficiency. Rt
Normally a number of electrolyzer cells are connected in series in a ′
Qcool = C cw (Tcwi − Tcwo ) = VAHX LMTD (14)
stack and some stacks are connected in parallel. The number of where Ct is overall thermal capacity of electrolyzer, T is the
such electrolyzer cells has an impact on the total hydrogen temperature in Kelvin, Q'gen is the internal heat generation, Q'loss is
production. The production rate is the total heat loss to the ambient, Q'cool is the auxiliary cooling
nc I demand, Rt is the overall thermal resistance of the electrolyzer, Ta
n H 2 = η F (9) is the ambient temperature, Ccw is the thermal capacity of the
zF cooling water, Tcwi is the temperature of the inlet cooling water,
where nc is the number of electrolyzer cells per stack. Tcwo is the temperature of outlet cooling water, VAHX is the
overall heat transfer coefficient and area product of heat
Energy efficiency is defined as the ratio of thermo neutral voltage exchanger and LMTD is the log mean temperature difference.
and the cell voltage as follows
Vtn The log mean temperature difference relates the temperature of
ηe = (10) electrolyte with that of inlet and outlet cooling water
V
Note that, the electrical inefficiencies result the unwanted heat (T − Tcwi ) − (T − Tcwo )
LMTD = (15)
generation in the electrolyzer which has an indirect influence on (T − Tcwi )
Faraday efficiency. The dynamic model describing electrochemical ln[ ]
property is shown in Figure 2. (T − Tcwo )

Figure 3: Thermal Simulation model.

713

Authorized licensed use limited to: PAKISTAN INST OF ENGINEERING AND APPLIED SCIENCES. Downloaded on December 11,2022 at 19:08:39 UTC from IEEE Xplore. Restrictions apply.
 As this study includes the dynamic modeling, it will be more viable When the electrolyzer is off, the cell voltage will be minimum
to analyze the differential equations and calculate temperature and will equal to reversible voltage (1.228 V) and when it is on, a
directly. But before that another unknown parameter, outlet cooling large amount of current will flow through the cells resulting
water temperature should be calculated first. increment in voltage (near about 1.54 V) even greater than the
VAHX thermo neutral cell voltage (1.48 V) as shown in Figure 4. The
Tcwo = Tcwi + (T − Tcwi )[1 − exp(− )] (16) cell voltage completely depends on current as it is the only
C cw variable term related to it [eq. (4) and eq. (7)]. The Faraday
Using equation (12), (13), (14) and (16) the overall thermal energy efficiency always represents the value less than 1 (see Figure 5).
balance from the basic equation (11) will become The electrolyzer becomes more efficient when the current is
stepped up. As the heat generation in the electrolyzer cell is due
dT to its inefficiency so when it remains in operation ηe decreases
+ AT − B = 0 (17)
dt below 1, in other case this value is greater than 1 as shown in
which has the solution of Figure 6. It is quite obvious that hydrogen production will be
increased as electrolyzer current increases. As shown in Figure 7
B B the hydrogen flow rate is 0.65 Nm3/sec per stack for an
T (t ) = (Tini − ) exp(− A × OT ) + (18)
electrolyzer stack current of 725A.
A A
where Tini is the initial temperature of the electrolyzer, OT is its
operating time, A and B are the arbitrary constants whose In thermal model it is easy to state that inlet cooling temperature
expressions are is less than the outlet cooling temperature, as a result the auxiliary
cooling is a negative quantity and the cooling rate has two steps
C cw
1 VA as shown in Figure 8. Thermal resistance is a small quantity
A= +[1 − exp(− HX )] (19)
leading to a heat loss (see Figure 9) while the rate of heat loss is
τ t Ct C cw constant with time. The electrolyzer temperature is quite stable
n VI (1 − η e ) Ta C cwTcwi VA with time (see Figure 10), the reason is the values of arbitrary
B= c + + [1 − exp(− HX )] constants A and B are very small compared to initial temperature.
Ct τt Ct C cw Hence it retains almost same temperature value for all conditions.
(20) Primarily V is less than Vtn causing a slight negative value of heat
where τt is thermal time constant and represented by the product of generation, but when the electrolyzer comes into operation as
Ct and Rt. shown in Figure 11, the heat generation changes its direction and
increases with time. Figure 12 shows the relationship between the
Empirically it has been found that there is an indirect relationship input current and power consumed by the electrolyzer. It is
between VA product and the electrolyzer current. Hence the 200kW while it is on.
relationship proposed by the author [1]
6. CONCLUSIONS
VAHX = hcond + hconv I (21)
where hcond and hconv are the parameters related to conduction heat Modeling and simulation of a 200kW alkaline type electrolyzer is
transfer and convection heat transfer respectively. Figure 3 presented in this paper. In order to make the study realistic, the
describes the dynamic thermal behavior of the electrolyzer. simulation time span has been selected as 24 hours. The input
current characteristic was chosen as a square wave. Simulation
5. SIMULATION RESULTS results are presented in figure 4-12. This study describes that the
more the current flows through the electrolyzer, the more the
The HySTATTM-A 1000D-30-25 model provided by Hydrogenics hydrogen will be produced. Simulation results show expected
Corporation is a 200 kW electrolyzer which is going to be installed system parameters and they match with the manufacture quoted
at Ramea. Its nominal hydrogen flow is 12-30 Nm3/h, output system parameters. The dynamic model presented here
pressure is 348 PSI and power requirement is 7.4 kWh/Nm3. There incorporates both electrochemical and thermal behavior of the
are two stacks and each stack contains 90 electrolyzer cells. electrolyzer. Presented alkaline type electrolyzer model can be
Although the dynamic model described here is a generalized integrated with the models of other renewable energy conversion
model, some input parameters have been chosen in such a way that systems to simulate dynamics of a hybrid power system
the output parameters match with the practical model.
Parameters of Practical Considered Results 7. ACKNOWLEDGEMENTS
electrolyzer Values Values Found
Power consumption 200 kW -- 200 kW This research is funded through National Science and
Engineering Research Council Wind Energy Strategic Network
Hydrogen Flow 12-30 -- 26
Rate Nm3/h Nm3/h (WESNet) grant. The authors would like to acknowledge Mr.
John Flynn of Newfoundland and Labrador Hydro and Rene J.
Cooling water <1.5 litre/h 0.6 litre/h --
Mandeville of Hydrogenics Corporation for their useful
consumption
discussions, suggestions and documents they provided us that
Inlet cooling water 10-30 °C 20 °C --
contributed this paper.
temperature

Table 1: Comparative listing of practical values, considered values

and simulation output values of different parameters.

714

Authorized licensed use limited to: PAKISTAN INST OF ENGINEERING AND APPLIED SCIENCES. Downloaded on December 11,2022 at 19:08:39 UTC from IEEE Xplore. Restrictions apply.
 Figure 4: Cell Voltage Figure 5: Faraday Efficiency Figure 6: Energy Efficiency

Figure 7: Hydrogen Production Figure 8: Auxiliary Cooling Figure 9: Heat Loss

Figure 10: Electrolyzer Temperature Figure 11: Heat Generation Figure 12: Input Current and Power
Consumption

8. REFERENCES

[1] Øystein Ulleberg, “Modeling of advanced alkaline [5] Gerhard Peharz, Frank Dimroth, UrsulaWittstadt, “Solar
electrolyzers: a system simulation approach”, International Journal hydrogen production bywater splitting with a conversion
of Hydrogen Energy, pp. 21-33, 28 (2003). efficiency of 18%”, International Journal of Hydrogen Energy,
pp. 3248 – 3252, 32 (2007).
[2] Haluk Görgün, “Dynamic modelling of a proton exchange
membrane (PEM) electrolyzer”, International Journal of Hydrogen [6] D.A. Bechrakis, E.J. McKeogh, P.D. Gallagher, “Simulation
Energy, pp. 29 – 38, 31 (2006). and operational assessment for a small autonomous wind–
hydrogen energy system”, Energy Conversion and Management,
[3] J.M. Vidueira, A. Contrerasa, T.N. Veziroglu, “PV autonomous pp. 46–59, 47 (2006).
installation to produce hydrogen via electrolysis, and its use in FC
buses”, International Journal of Hydrogen Energy, pp. 927– 937, [7] Kodjo Agbossou, Mohanlal Kolhe, Jean Hamelin, and Tapan
28 (2003). K. Bose, ” Performance of a Stand-Alone Renewable Energy
System Based on Energy Storage as Hydrogen”, IEEE
[4] Keliang Zhou, J.A. Ferreira, S.W.H. de Haan, “Optimal energy TRANSACTIONS ON ENERGY CONVERSION, pp. 633-640,
management strategy and system sizing method for stand-alone VOL. 19, NO. 3, SEPTEMBER 2004.
photovoltaic-hydrogen systems”, INTERNATIONAL JOURNAL
OF HYDROGEN ENERGY, pp. 477 – 489, 33 (2008).

715

Authorized licensed use limited to: PAKISTAN INST OF ENGINEERING AND APPLIED SCIENCES. Downloaded on December 11,2022 at 19:08:39 UTC from IEEE Xplore. Restrictions apply.