Journal of Energy Storage 26 (2019) 101024

Contents lists available at ScienceDirect

Journal of Energy Storage

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

Comparative study of different storage bed designs of a solid-state hydrogen T

tank
Mohamed Sakreddine Manaia,b, Mikel Leturiaa, Carsten Pohlmannb, Jorn Oubrahamb,
⁎
Stéphane Motteleta, Michael Levyb, Khashayar Saleha,
a
Sorbonne Universités, Université de Technologie de Compiègne, Integrated Transformations of Renewable Matter Laboratory (EA TIMR 4297 UTC-ESCOM), rue du Dr
Schweitzer, 60200 Compiègne, France
b
Aaqius, 12 rue Vivienne, 75002 Paris, France

A R T I C LE I N FO A B S T R A C T

Keywords: This work discusses the influence of different metal hydride storage bed configurations. The objective was to
Solid-state hydrogen storage design and optimize a solid-state hydrogen storage for a nonpolluting mobility.
Numerical MATLAB based simulation A study of the absorption and desorption dynamics of a loose powder bed was performed first, followed by
AB2 three different storage bed configurations: compacted Ti-Mn alloy powder, alternated Ti-Mn alloy compacts with
Heat and mass transfer
stainless steel fins and compacted [Ti-Mn alloy/Stainless steel] powder mixture. A numerical model was de-
Hydrogen storage bed design
veloped to simulate the heat transfer and the hydrogen absorption and desorption rates.
The alternation and compact mixture configurations gave better heat transfer efficiencies, absorption and
desorption rates and increased hydrogen storage densities. Indeed, an efficient heat transfer (between the tank
and its surrounding fluid), a tailored porosity of the metal hydride storage bed and the addition of high thermal
conductivity materials allowed the overall storage performance to be improved. Thus, the required time for
loading/unloading hydrogen was reduced drastically. The alternation configuration would offer the additional
advantage of a simple, inexpensive and efficient recycling procedure.

1. Introduction storage [8,9].

The storage of hydrogen in gaseous form is the most widely used
The irreversible depletion of fossil fuels opens a debate between technology today. This technology is mastered up to 700 bar [10]. Li-
scientists, industry and civil society. This debate highlights the urgent quid storage requires hydrogen cooling to extremely low temperatures:
need to search for alternative and renewable sources of energy. Not 5 K (− 268.15 ∘C) [11]. In this context, solid-state hydrogen storage
only does the depletion of fossil fuels push us to think about replacing offers several advantages: reducing the risks related to pressurized
them with renewable energy sources, but also the environmental crisis storage, no need of ultra-cool temperature management, flexibility of
will have serious consequences on the way of life of human beings use, safety, compact, efficiency, just to name a few [12]. The storage of
[1,2]. hydrogen in a solid media (or solid-state storage) includes storing hy-
It is therefore urgent to reduce the consumption of fossil energy and drogen either by absorption in a carrier material to form a hydride, or
it is necessary to accelerate research of other non-polluting and re- by surface adsorption [13,14].
newable energy sources and carriers. Renewable resources such as wind This study is focused on the storage of hydrogen in form of metal
and solar, converted to electricity can play an important role in the hydrides. The reaction between hydrogen and the metal alloy can be
energy transition [3,4]. Electricity storage is a solution to limit carbon described as follows:
emissions by enabling more use of intermittent renewable energy. M + x H2 ↔ MH2x + Heat (1)
Nowadays, solutions are available such as electrochemical accumula-
tors, capacitors, chemical storage, as well as thermal storage. Among The reaction produces a metal hydride with heat release (exo-
these solutions, hydrogen storage (chemical storage) offers an inter- thermic reaction). In metal hydrides, hydrogen occupies the interstitial
esting potential [5–7]. In fact, hydrogen storage is generally realized via sites in the crystal structure of the metal. This interaction causes an
three different technologies: gas storage, liquid storage and “solid” expansion of the structure of the alloy. Generally, a hydrogen loading at

⁎
Corresponding author.
E-mail address: [email protected] (K. Saleh).

https://doi.org/10.1016/j.est.2019.101024
Received 19 July 2019; Received in revised form 16 October 2019; Accepted 16 October 2019
2352-152X/ © 2019 Published by Elsevier Ltd.
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

Nomenclature V gas velocity, m/s

Ca absorption reaction rate coefficient, 1/s Greek

Cd desorption reaction rate coefficient, 1/s
Cp specific heat capacity at constant pressure, J/(kg.K) ΔH enthalpy of reaction per mole, J/kg
Ea activation energy, J/mol ɛ porosity, –
h heat transfer coefficient, W/(m2 . K) ρ mass density, kg/m3
K permeability, m2 μg dynamic viscosity, Pa .s
k thermal conductivity, W/(m.K)
M molar mass, g/mol Subscripts
ṁ reaction rate, kg/(s.m3)
Pg hydrogen pressure, bar abs absorption
Patm standard pressure, 1.01325 bar axi axial
R universal gas constant, 8.3145 J/(mol.K) des desorption
T temperature, K eff effective (bulk)
Tf temperature of the cooling/heating fluid, K eq equilibrium
t time, s H hydrogen
t90%_abs time to reach 90% of reacted fraction, s i index
t90%_des time to reach 10% of reacted fraction, s

a pressure comparable to the operating pressure of an electrolyzer transfer phenomena. The objective of this work was to compare dif-
(about 10 bar) [15,16] and an unloading at the pressure of a fuel cell ferent heat transfer enhancement techniques in order to optimize the
(about 2 bar) [17,18] are sought. hydrogen storage tank dynamics. For this purpose, different config-
A large number of applications of metal hydride tanks has been urations of the AB2 Ti-Mn alloy storage bed were discussed and ana-
reported in the literature. Bellosta von Colbe et al. [19] proposed a lyzed.
summary of metal hydrides applications for hydrogen storage and This work is based on the results published by Herbrig et al. [36].
compression. For stationary applications, the weight of the tank does The material chosen for the simulation of the absorption and desorption
not represent a technological barrier. Consequently, the tank weight tests is a titanium-manganese alloy with the following chemical formula
only depends on the hydrogen amount which is needed to be stored. For Ti0.95Zr0.05Mn1.48Fe0.08Al0.01 (AB2 type). This Ti-Mn alloy is also com-
off-grid systems (stand-alone power systems), metal hydride systems mercially called Hydralloy C5 [37–40]. It can reach a gravimetric hy-
can be a support for the integration of renewable energy sources (wind drogen storage capacity of about 1.8 wt% at room temperature [36].
energy, photovoltaic energy, hydroelectric energy, etc.). Bielmann et al. The large volumetric capacity of this metal hydride is also an asset for
[20] proposed an energy system for isolated sites and remote areas. This automotive applications regarding the applicable volume constraints. A
energy system included an AB5 type metal hydride and a Li-ion battery. 2D axisymmetric numerical model was developed on MATLAB to study
Song et al. [21] tested a 12 kg hydrogen storage tank for a mobile light the limiting mechanism, which was either the external heat transfer
tower used on construction sites. Finally, metal hydride tanks could also (convective mechanism between the tank and the coolant fluid) or the
play a role in the development of smart-grid energy storage [22–26] internal heat transfer (conductive mechanism in the metal hydride
and as a compressor [27–29]. bed). For this purpose, different simulations were performed:
Heat transfer management is a major factor in order to control the
hydrogen loading and unloading times of a storage tank. Afzal et al. - First, the model was validated by comparison with the experimental
[30,31] carried out a study that groups together different heat transfer results published by Herbrig et al. [36];
techniques: improvement of the thermal conductivity by inclusion of - Second, the influence of the external heat transfer was studied, by
ENG in the metal hydride bed; increase of the heat transfer area by modifying the heat transfer coefficient h (convective heat transfer
addition of fins and cooling tubes (to improve internal heat transfer); between the tank walls and the coolant fluid). Hydrogen absorption
improvement of the external heat transfer by increasing the mass flow and desorption simulations were performed on the reference con-
rate of the heat transfer fluid, etc. Bhogilla et al. [32] designed a cy- figuration (“Reference”), which corresponds to a loose metal hy-
lindrical tank with a coil-shaped heat exchanger inside. This allows heat dride powder bed;
transfer from the system directly to the hydride bed without, passing - Third, different internal heat transfer enhancement techniques were
through the outer wall (elimination of the external heat transfer re- simulated (compaction and/or addition of a high thermal con-
sistance). The use of heat transfer tubes inside the tank was also studied ductivity material). The reference configuration (“Reference”) was
by Ranong et al. [33] to show the effect of the number of tubes on heat compared to 3 metal hydride bed configurations: “Compact”,
transfer and thus on the time required for loading and unloading hy- “Alternation” and “Compact mixture”. The comparison was based
drogen. Mellouli et al. [34] and Dhaou et al. [35] studied the effect of a on the absorption rate (apparent kinetics) as well as the total hy-
coil heat exchanger but also analyzed the effect of adding aluminum drogen storage capacity;
fins. This last technique allows a 50% performance increase compared - Finally, the Alternation configuration (alternation of metal hydride
to a basic tank configuration. In these studies, heat transfer enhance- compacts and stainless steel fins) was found to offer a good com-
ment techniques were studied experimentally and numerically. The promise between the total hydrogen storage capacity, the absorption
comparison of these techniques is required to choose and optimize the rate and ease of implementation. This configuration was used to
appropriate technique for a given storage system. analyze the influence of the 2 main operating parameters, i.e.,
This study focuses on the modeling and comparison of different temperature and pressure.
configurations of an AB2 Ti-Mn alloy storage bed. The hydrogen ab-
sorption (desorption) reaction of the AB2 metal hydride alloy is fast and It should be noted that the simulation of multi-material configura-
the temperature can increase (decrease) rapidly at the beginning of the tions for heat transfer enhancement (Alternation and Compact mixture)
reaction. Therefore, the apparent kinetics is frequently limited by heat represents a novelty in this work. It shows the role of a high thermal

2
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

conductivity materials on the improvement of the hydrogen storage (ρs Cps )e = [βv ρsteel Cpsteel + (1 − βv ) ρMH CpMH ] (4)
performance.
where βv is the volumetric fraction of stainless steel in the storage bed.
2. Model description and validation The effective thermal conductivity can be calculated as:

ke = [εk g + (1 − ε ) ks] (5)

2.1. Metal hydride tank geometry
ks = [βv ksteel + (1 − βv ) kMH ] (6)
The tank considered in this work (Fig. 1) corresponds to the con-
figuration used by Herbrig et al. [36] to study a hydrogen storage The hydrogen gas density is calculated applying the ideal gas law:
system based on a graphite-metal hydride composite. The experimental
Pg Mg
results published by Herbrig et al. [36] are used to validate the 2D ρg =
RT (7)
axisymmetric numerical model developed on MATLAB.
A cylindrical tank with an internal diameter of 40 mm is considered. where Pg is the hydrogen gas pressure, Mg is the molar mass of the gas
The diameter of the internal chimney is 6 mm. Hydrogen enters through and R is the universal gas constant.
this chimney to diffuse into the metal hydride storage bed. The tank The gas velocity in the bed is calculated using Darcy's law:
wall is 8 mm thick and is made of stainless steel. The internal height of
the storage bed is 130 mm. → K→
V = − ∇P
μg (8)
2.2. Governing equations
with:
In the present model, geometric, thermal, thermodynamic and ki- T 0.68
netic variables are considered. This simulation takes the external soli- μg = 9.05 × 10−6 × ⎛ ⎞
⎝ 293 ⎠ (9)
citations, such as the variation of the hydrogen pressure and the ex-
ternal heating or cooling of the tank into account. Different where μg is the dynamic viscosity of hydrogen and K is the permeability
mathematical models detailed in the literature have been developed to of the metal hydride storage bed.
describe the behavior of metal hydride beds during hydrogen absorp- The continuity equation is expressed by:
tion and desorption processes [41–46]. Indeed, the mass and heat →
∂
transfer simulation are essentially based on the mass and energy con- ε ρ + div(ρg V ) = −m˙
∂t g (10)
servation laws, which can be expressed by partial differential equations
(PDEs). After combining continuity Eq. (10), Darcy's law (8) and ideal gas
The assumptions considered in the present model are the following: law (7), Eq. (11) can be obtained:

Mg ⎛ ∂Pg Pg ∂T ⎞ K →
- The model is axisymmetric in a cylindrical coordinate system (r, z). ε ⎜ − =
⎟ div (ρg ∇ P ) − m˙
- At any location, the temperatures of the solid materials and the gas R T ⎝ ∂t T ∂t ⎠ μg (11)
are assumed to be equal (local thermal equilibrium).
- The convective heat transfer (trough the porous medium) is negli- The solid mass balance is given by:
gible. ∂ρs
- Hydrogen behaves like an ideal gas. (1 − ε ) = m˙
∂t (12)
- The expansion (contraction) of the metal hydride during hydro-
genation (dehydrogenation) is not taken into account. where ṁ corresponds to the source term of the reaction and ρs is the
- The porosity and thermal conductivity are uniform and their var- density of the metal hydride. The source term ṁ takes a different ex-
iation with the absorption and desorption processes are neglected. pression for absorption and desorption.
The amount of hydrogen absorbed by the metal over time is directly
It is important to point out that the model considers different phy- related to the hydrogenation kinetics, expressed as:
sical parameters for each material involved (metal hydride storage bed,
tank walls, etc.). For the ease of understanding, the equations given
below concern only the metal hydride storage bed.
Taking the hypothesis of local thermal equilibrium between the gas
and the solid materials, a single equation of the energy balance is ap-
plicable:
∂T 1 ∂ ⎛ ∂T ⎞ ∂ ⎛ ∂T ⎞
(ρCp)e − rke − ke = −m˙ [ΔH + T (Cpg − Cps )]
∂t r ∂r ⎝ ∂r ⎠ ∂z ⎝ ∂z ⎠
(2)
where T is the temperature, (ρCp)e is the effective thermal capacity, ke is
the effective thermal conductivity, ṁ is the source/sink term, ΔH is the
enthalpy of reaction, Cpg and Cps are respectively the thermal capacities
of the gas and the solid.
Since the hydrogenation and dehydrogenation reactions within the
hydride bed are exothermic (ΔH < 0) and endothermic (ΔH > 0), re-
spectively, heat is produced or consumed. Therefore, a source/sink
term related to absorption/desorption ṁ is added to the energy balance
equation.
The effective heat capacity can be expressed as: Fig. 1. Simplified schematic diagram of the metal hydride tank used for the
simulations (Points A and B refer to the temperature probe positions, ‘core’ and
(ρCp)e = [ερg Cpg + (1 − ε ) ρs Cps] (3) ‘border’, respectively).

3
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

E Pg ⎞ - Initial conditions:
m˙ abs = Ca exp ⎛− a ⎞ ln ⎜⎛ ⎟ (ρsat − ρs )
⎝ RT ⎠ ⎝ Peq _ abs ⎠ (13) T (r , z , t = 0) = Tinitial (17)
where Ca is the absorption kinetic constant (which depends on the
material used for hydrogen storage), Ea is the activation energy for - Boundary conditions:
absorption, ρsat is the density of the saturated metal hydride (i.e., when
dT (r = R, z , t ) h
it has absorbed all the hydrogen that may be reversibly absorbed at a = (Tf − T (r = R, z , t ))
given hydrogen pressure and temperature) and Peq _ abs is the equilibrium dr ke (18)
pressure for absorption.
Similarly, the amount of hydrogen desorbed by the metal over time
is directly related to the metal desorption kinetics, expressed by: dT (r = 0, z , t )
=0
dr (19)
E Pg − Peq _ des
m˙ des = Cd exp ⎛− d ⎞ (ρs − ρemp )
⎝ RT ⎠ Peq _ des (14)

where Cd is the desorption kinetic constant, Ed is the activation energy where h is the heat transfer coefficient and Tf is the temperature of the
for desorption, ρemp is the density of the metal hydride when it has heating/cooling fluid surrounding the tank.
desorbed all the hydrogen which can be reversibly desorbed and Peq _ des
is the equilibrium pressure for desorption. It should be noted that the 2.3. Model validation
equilibrium pressure for absorption and desorption are given by Eqs.
(21) and (22) in the study of Herbrig et al. [36]. The same parameters and simulation conditions as those defined by
The hydrogen content is expressed as: Herbrig et al. [36] were used to validate the mathematical model de-
ρs − ρemp veloped in the present work. A hydrogen loading simulation was re-
wt = produced and compared to the experimental results published by Her-
ρs (15)
brig et al. [36]. The cooling fluid temperature was set to 50 ∘C and the
The conversion (or reacted fraction) is defined as: hydrogen pressure was 40 bar (Fig. 2). The key parameters of the si-
wt − wt0 mulation are shown in the Table 1.
Conversion = As shown in Fig. 2, the results of the simulation are in good
wtmax − wt0 (16)
agreement with those given by Herbrig et al. [36], with an acceptable
where wt0 is the initial hydrogen content (before absorption) and wtmax deviation. The same temperature evolution, which reaches a maximum
is the hydrogen content of the metal hydride after complete absorption temperature (since the reaction is exothermic) to finally decrease until
(i.e., for the fully hydrogenated metal hydride). reaching the cooling temperature, is observed. At the end of the reac-
The hydrogen is introduced axially inside the hydride bed through a tion (experimental results), there is a flexion point which is also visible
porous filter. The initial and boundary conditions can be expressed as in the simulation results. This point is the consequence of the end of the
follows: reaction. The simulated temperature is generally higher than the

Fig. 2. Simulation results for the temperature (point A in Fig. 1) and conversion as a function of time during hydrogenation at 40 bar and 50 °C, in comparison with
experimental results of Herbrig et al. [36].

4
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

Table 1 3.1. Configuration 1: “Reference”

Simulation parameters.
Parameter Value References The hydrogenation of a metal hydride powder bed is simulated. The
hydrogen inlet pressure is 40 bar and a cooling temperature of 50 ∘C is
CpMH 500 J/(kg K) [36] set. The dehydrogenation simulation is performed for a heating tem-
Cpsteel 460 J/(kg K) [36]
perature of 50 ∘C at 1 bar backpressure (the pressure and temperature
Ea 500 J/mol [38]
Ca 6 1061/s [38]
values are the same as those used for the validation of the model). Metal
ΔHabs − 11.15 106 J/kg [36] hydride parameters for the simulation are given in Table 1. The hy-
ΔHdes 14.2 106 J/kg [36] drogen storage bed of the Reference configuration is shown in Fig. 3.a.
kMH 1 W/(m K) [36]
ksteel 15 W/(m K) [36]
MMH 0.154 kg/mol [36] 3.2. Configuration 2: “Compact”
MH2 0.02 kg/mol [36]
ɛ 0.31 m3/m3 [36] By compacting the powder, it is possible to reduce the porosity and,
K 8.1 10−15m2 [36] therefore, to improve the volumetric storage capacity as well as the
ρemp 6100 kg/m3 [36] effective thermal conductivity (Eq. (5)), i.e., the internal heat transfer.
ρsat 6212 kg/m3 Calculated here for wtmax=1,8 wt%
ρsteel 7900 kg/m3 [36]
The total apparent volume of metal hydride (i.e., internal tank volume)
h 500 W/(m2 K) [36] is the same as in the first configuration (Reference). The values for
porosity (0.31), permeability (8.1 10−15 m2 ) and thermal conductivity of
the metal hydride bed (1.22 W/(m.K)) are chosen according to Herbrig
temperature observed during the hydrogenation experiment. Despite et al. [36]. The hydrogen storage bed of the Compact configuration is
the lack of a complete description of the operating conditions of the shown in Fig. 3b.
experiment carried out by Herbrig et al. [36], the simulation results are
still similar to the experimental results. Among the missing information 3.3. Configuration 3: “Alternation”
is the external heat transfer coefficient (h). Also, in this simulation, the
tank is supposed to be immersed in an isothermal bath with constant In the Alternation configuration (Fig. 3c), the advantage of com-
cooling/heating temperature. paction (increase of the hydrogen volumetric capacity) is kept. Fur-
In the following sections of this paper, this model will be used from thermore, to improve internal heat transfer, a material with a high
a comparative perspective. It will allow the investigation of the influ- thermal conductivity (generally much higher than the metal hydride
ence of external and internal heat transfers. The objective is to improve compact) is placed between each compact. The storage bed will
the system performance characteristics (i.e., total hydrogen storage therefore consist of an alternation of metal hydride compacts and high
capacity and absorption/desorption rates). thermal conductivity material zones (Fig. 3c). Stainless steel is chosen
for its low price and high abundance. The thermal conductivity of
3. Description of the metal hydride bed configurations stainless steel is 15 times greater than the thermal conductivity of the
metal hydride.
The absorption reaction is fast, and the temperature of the reaction The storage bed is in fact composed of 6 metal hydride compacts.
increases rapidly during this reaction. The apparent kinetics of the tank Five fins are placed between the compacts. This configuration results in
system is therefore limited by the reaction equilibrium. The Reference a mass ratio of 13 wt% of stainless steel in the storage bed.
configuration will be used to investigate the influence of the external
heat transfer (convective heat transfer between the tank walls and the 3.4. Configuration 4: “Compact mixture”
coolant fluid). The other configurations (Compact, Alternation and
Compact mixture) will be studied to show the effect of the internal In the Compact mixture configuration Fig. 3d), the compacts are
heat transfer (conductive heat transfer in the metal hydride bed) on the composed of a mixture of metal hydride powder and 13 wt% of stainless
system dynamics (absorption rate). To improve the internal heat steel powder. The mass ratio of 13 wt% of stainless steel is chosen to
transfer, two possible options used separately or together will be con- allow the comparison between configurations 3 and 4. The considered
sidered: powder compaction and/or addition of a high thermal con- compacts exhibit a porosity of 0.31 (note that the compact mixture can
ductivity material into the hydride bed (e.g., insertion of fins). reach the same porosity by applying a given pressure, different from the

Fig. 3. Schematic diagram of the different configura-

tions of the metal hydride bed: (a) Ti-Mn alloy powder
bed (1st configuration: “reference”), (b) Ti-Mn alloy
compact bed (2nd configuration: “Compact”), (c)
Alternation of stainless steel fins and Ti-Mn alloy
compacts (3rd configuration: “Alternation”) and (d)
Compact of a mixture of stainless steel and metal hy-
dride powders (4th configuration: “Compact mix-
ture”).

5
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

compacting pressure of the compact of configurations 2 and 3) and a comparison was based on the absorption rate (apparent kinetics) as
permeability of 7.5 10−15m2 [36]. The thermal conductivity of this well as the total hydrogen storage capacity;
compacted mixture is calculated according to Eqs. (5) and ((6). - Finally, the Alternation configuration was found to offer a good
The loading and unloading conditions are the same for all simula- compromise between the total hydrogen storage capacity, the ab-
tions: the loading (unloading) is performed under isothermal condition sorption rate and ease of implementation. This configuration was
at 50∘C and an inlet pressure of 40 bar (1 bar). The heat transfer coef- used to analyze the influence of the 2 main operating parameters,
ficient is h = 500 W/(m2. K) [36]. i.e., temperature and pressure.

4. Results and discussions 4.1. Influence of the external heat transfer

One of the major controlling parameters of the solid hydrogen sto- The influence of the heat transfer coefficient h between the storage
rage process is the heat transfer in/to the metal hydride bed. Indeed, tank and the temperature-controlled bath (surrounding the tank) is
from the thermodynamics point of view, the hydrogenation reaction presented in this section. The range of values was chosen between 100
can be limited without improvement of the heat transfer. To study and 3000 W/(m2 K) [36]. For comparison, the heat transfer coefficient
which heat transfer resistance is limiting the apparent kinetics, various of an air-cooled system under natural convection is between 2 and
simulations on different configurations of the metal hydride bed were 25 W/(m2 K), which can be improved up to 25 to 250 W/(m2 K) under
performed: forced convection. Using oil as a cooling liquid, the effect is even more
pronounced with 50 to 2000 W/(m2 K) under forced convection [47]. In
- First, the influence of the external heat transfer was studied, by a nutshell, the improvement of the external heat transfer can be done by
modifying the heat transfer coefficient h (convective heat transfer an appropriate choice of the applied cooling / heating method (natural
between the tank walls and the coolant fluid). Hydrogen absorption or forced) and the used cooling fluid (air, water, oil, etc.).
and desorption simulations were performed on the Reference con- Fig. 4a shows the evolution of the temperature as a function of time
figuration; (P = 40 bar , Tf = 50 ∘C ). At first, the temperature increases rapidly
- Second, the different internal heat transfer enhancement techniques to reach a maximum value, followed by a slow decrease approaching
were simulated. The Reference configuration was compared to the the cooling temperature Tf. The temperature increases as a result of the
Compact, Alternation and Compact mixture configurations. The exothermicity of the reaction. The cooling of the tank allows the

Fig. 4. Simulation results of the Reference configuration, (a) Temperature at the core of the metal hydride bed (point A in Fig. 1), (b) Reacted fraction during
hydrogenation at 40 bar and 50 °C, (c) Temperature at the core of the metal hydride bed, (d) Reacted fraction during dehydrogenation at 1 bar and 50 °C for different
heat transfer coefficient values.

6
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

temperature lowering of the metal hydride powder bed. Fig. 4b shows Compact, Alternation and Compact mixture configurations are stu-
the evolution of the conversion (or reacted fraction) as a function of died (with h = 500 W/(m2 K) , [36]). The simulation results are pre-
time. By increasing the value of the heat transfer coefficient, the reac- sented in Fig. 6 and summarized in Table 2 (t90% _ abs , t90% _ des and total
tion time for both absorption and desorption decreases. Interestingly, amount of hydrogen stored).
the external heat transfer coefficient value has a negligible influence on For mobile applications, it is important that the hydrogen loading/
the magnitude of the peak temperature. However, for large heat unloading is done as quickly as possible. In this part of the work, the
transfer coefficients (h), the return of the hydride bed temperature to management of the internal heat transfer was studied for the loading
the cooling temperature Tf is faster (Fig. 4a), which leads to a faster step. The same conclusions would also be valid for the unloading step.
hydrogen absorption (Fig. 4b). These results also indicate that the in-
ternal heat transfer is a limiting factor.
The time required to reach 90% reacted fraction (conversion) 4.2.1. Configuration 2: “Compact”
during absorption and 10% reacted fraction (conversion) during deso- Powder compaction allows a higher effective thermal conductivity
rption can be used to characterize the overall system dynamics (noted and also substantially increases the amount of hydrogen that can be
t90% _abs and t90% _des for absorption and desorption, respectively). These stored in the tank. Indeed, compacting the metal hydride powder leads
2 characteristic times obtained from the simulation results are pre- to a higher apparent density of the metal hydride bed. Consequently,
sented in Fig. 5. The results show that there is a decrease in the loading more metal hydride can be placed in the same tank volume, which
time t90% _abs when increasing the heat transfer coefficient: for example, increases the amount of hydrogen that can be stored in the tank.
668 s for a coefficient of h = 100 W/(m2 K) and 344 s for The simulation results (Fig. 6) show that by compacting the powder
h = 3000 W/(m2 K) . For desorption (Fig. 4c and d), the increase of the and reducing the porosity of the metal hydride bed, the peak tem-
heat transfer coefficient has a smaller effect on the unloading time t90% perature during hydrogen loading is less important (from 94 ∘C for the
2
_des. for example, 372 s for a coefficient of h = 100 W/(m K) and 246 s Reference configuration to 88 ∘C for the Compact configuration). The
2
for h = 3000 W/(m K) . characteristic absorption time t90% _ abs is increased, from 404 s for the
Between h = 100 W/(m2 K) and h = 500 W/(m2 K) , the increase of Reference configuration to 805 s for the Compact configuration. This
the heat transfer coefficient leads to a faster apparent kinetics. Above a can be explained by the higher total metal hydride mass (Compact
value of 500 W/(m2 K), no significant influence of the external heat configuration), which increases the amount of hydrogen that can be
transfer on the apparent kinetics is observed. Even with a 6-fold in- stored in the tank and provokes a longer loading time.
crease of this value, the time required to reach t90% _abs is not sig- For the same quantity of hydrogen stored, the second configuration
nificantly reduced. It passes from t90% _ abs = 405 s for a heat transfer is more efficient from the point of view of apparent kinetics. For ex-
coefficient h = 500 W/(m2 K) to t90% _ abs = 344 s for a heat transfer ample, for the Reference configuration, 4.53 g of hydrogen are stored
coefficient h = 3000 W/(m2 K) . in the tank at t90% _ abs (= 404 s ). For the Compact configuration, the
Overall, these results show that above a value of 500 W/(m2 K), the tank only needs 44 s to store the same amount of hydrogen.
internal heat transfer becomes the rate limiting step of the absorption The effective thermal conductivity of the hydride bed is increased
and desorption processes. This conclusion can be explained by the poor by 22%, from 1 W/(m K) for the Reference configuration (loose
thermal conductivity of the hydride powder. Therefore, the improve- powder) to 1.22 W/(m.K) for the Compact configuration (compacted
ment of the internal heat transfer should be investigated by examining powder) [36]. At the same time, the storage capacity increases from
different configurations corresponding to different effective thermal 5.05 g to 11.55 g (+130%). This necessitates to withdraw much more
conductivities of the hydride bed. heat during the reaction. The increase of the effective thermal con-
ductivity of the hydride bed alone is not sufficient to compensate for the
increase of the heat released by the hydrogenation reaction. For this
4.2. Management of the internal heat transfer
reason, the addition of fins in the metal hydride bed (configuration 3:
Alternation) is considered in the following section, to further improve
The reduction of the porosity (compaction) of the storage bed and/
the internal heat transfer.
or addition of a material with high thermal conductivity is suitable to
tailor the internal heat transfer characteristics. For this purpose, the

Fig. 5. Hydrogenation and dehydrogenation times (t90%) of the Ti-Mn alloy tank system for different heat transfer coefficient values.

7
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

Fig. 6. Simulation results for the different storage bed configurations, (a) Temperature at the core of the metal hydride bed (point A in Fig. 1), (b) Reacted fraction
during hydrogenation at 40 bar and 50 °C.

Table 2
Characteristic times (t90%_abs, t90%_des) and total amount of hydrogen stored for the different tank configurations.
Configuration t90%_abs [s] hydrogenation t90%_des [s] dehydrogenation mMH [g] mbed [g] mH2_tot [g]

1 404 280 298 298 5.05

2 805 645 682 682 11.55
3 535 440 630 725 10.67
4 312 272 630 725 10.67

4.2.2. Configuration 3: “Alternation” efficient from the point of view of apparent kinetics. In summary, the
The effect of the additional fins on the absorption behavior of the Alternation configuration allows better heat transfer characteristics,
metal hydride bed was tested by performing absorption simulations. As higher absorption rates (hydrogen loading rates) as well as a larger
shown in Fig. 6, the addition of fins in the “Alternation” configuration hydrogen storage capacity.
allows a faster heat dissipation. Indeed, it is observed that the peak
temperature measured during absorption is significantly reduced. In-
itially, the absorption rate is higher than the heat transfer rate, and as a 4.2.3. Configuration 4: “Compact mixture”
For the Compact mixture configuration, the absorption simulation
result, most of the energy released due to hydrogen absorption is ab-
sorbed by the hydride bed itself, which leads to a sudden rise in the bed (Fig. 6) shows a temperature profile which is broadly similar to the
other configurations (fast increase until a peak temperature, followed
temperature. Then, the bed temperature gradually decreases as the
energy released by the reaction reaches the tank walls (external heat by a slow decrease approaching the cooling temperature). However, the
temperature level is different. In this case, the peak temperature is si-
transfer between the tank walls and coolant fluid). The addition of fins
involves the addition of heat transfer surface area into the metal hy- milar to the one observed for the Alternation configuration, i.e.,
around 80 ∘C. The homogeneous distribution of the steel powder in the
dride system. It improves the effective heat transfer characteristics by
providing high thermal conductivity paths. It has a significant impact metal hydride compacts allows an important increase of the effective
thermal conductivity compared to the thermal conductivity of the metal
upon the performance of the reactor.
The second and the third configurations (Compact and Alternation hydride compacts alone (i.e., without addition of stainless steel
configurations) can store approximately the same amount of hydrogen powder). The cooling of the compact following the increase of the
(around 11 g). temperature is much faster than for any other configuration. Thus, 90%
However, the characteristic absorption time t90% _ abs is significantly conversion is achieved in 312 s (t90% _ abs ). The addition of stainless steel
powder improves the performance of the reactor, leading to better in-
reduced, from 805 s for the Compact configuration to 535 s for the
Alternation configuration. Therefore, this third configuration is more ternal heat transfer and thereby reduces the bed temperature as ob-
served in Fig. 6.a. Also, the stainless steel powder leads to a more even

8
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

temperature distribution along the metal hydride bed. pressure.

Among the three configurations studied to improve the internal heat
transfer, the Compact mixture configuration has the strongest impact 4.3. Influence of the hydrogen pressure and cooling fluid temperature
on increasing the apparent kinetics of the system. For the same tank (Alternation configuration)
volume, it allows a high hydrogen storage capacity with a fast ab-
sorption rate. Various simulations have been performed with the Alternation
Table 2 summarizes the results obtained for the 4 configurations. In configuration in order to investigate the influence of the hydrogen
summary, the Compact configuration allows a larger hydrogen storage pressure and the cooling fluid temperature. The reference operating
capacity but only a slight improvement of the internal heat transfer, conditions used in the previous sections of this paper are 40 bar for the
compared to the Reference configuration. By adding fins to the com- hydrogen pressure (Pg) and 50 ∘C for the cooling fluid temperature (Tf)
pacted metal hydride powder (Compact configuration), the hydrogen and for the initial temperature (Tinitial) of the storage bed. The corre-
loading rate is greatly improved and the amount of hydrogen stored is sponding results for the Alternation configuration were presented in
significantly increased. Of all the tested configurations, the Compact Fig. 6. Consequently, the following operating conditions have been
mixture configuration provides the fastest loading rate, with a high tested:
hydrogen storage capacity. However, it should be noted that compared
to the Alternation configuration (metal hydride compacts + stainless - Hydrogen pressure: 25 bar, 35 bar, 45 bar, 55 bar, with a cooling
steel fins), the Compact mixture configuration (metal hy- fluid temperature (Tf) and an initial temperature (Tinitial) of 50 ∘C;
dride + stainless steel compacted powder mixture) would be difficult to - Initial and cooling fluid temperature: 35 ∘C, 45 ∘C, 55 ∘C, 65 ∘C, with
separate and recycle. Indeed, a complex and expensive procedure a hydrogen pressure (Pg) of 40 bar.
would be required to separate the two powders at the end of the storage
tank life, resulting in a difficult recycling process. Overall, the concept The obtained results are presented in Fig. 7. Also, it should be noted
of alternating compacted storage material powders with high thermal that when the pressure and temperature are changed, the amount of
conductivity fins (Alternation configuration) gives the most promising hydrogen absorbed at equilibrium varies (absorption isotherms are
results and offers a good compromise between the total hydrogen sto- described by equations (21) and (22) in the study of Herbrig et al. [36]).
rage capacity, absorption rate and ease of implementation. This con- The amount of hydrogen absorbed at equilibrium (i.e., hydrogen sto-
figuration will be investigated in the following section, by analyzing the rage capacity) for all the tested operating conditions are summarized in
influence of the 2 main operating parameters, i.e., temperature and Fig. 8.

Fig. 7. Simulation results for the Alternation configuration, (a) Temperature and (b) Conversion as a function of time for different hydrogen pressures, (c)
Temperature and (d) Conversion as a function of time for different cooling fluid temperatures.

9
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

Fig. 8. Hydrogen storage capacity for the Alternation configuration at various operating conditions.

Regarding the influence of the hydrogen pressure (Pg), the simula- Finally, it should be noted that the reference operating conditions
tion results show that: (40 bar and 50 ∘C) chosen in this work are reasonable as they allow a
high hydrogen capacity (Fig. 8).
- The hydrogen storage capacity (Fig. 8) is larger when the pressure
increases (in agreement with the absorption isotherms described by
Eqs. (21) and (22) in the study of Herbrig et al. [36]). This storage 5. Conclusion
capacity increase is significant between 25 and 35 bar, but is small
above 35 bar; A MATLAB based model was developed to simulate the hydro-
- The hydrogen absorption rate (Fig. 7b) is faster at high pressures. genation and dehydrogenation of an AB2 metal hydride contained in a
This result is in agreement with the hydrogenation kinetics hydrogen storage tank. This 2D axisymmetric model allowed the si-
(Eq. (13)) which shows that ṁabs increases for a larger pressure (Pg); mulation of hydrogen absorption and desorption for various tank con-
- The peak temperature observed at the beginning of the hydrogen figurations, corresponding to different heat transfer enhancement
absorption is also larger at high pressures (Fig. 7a). As the hydrogen techniques (powder compaction and/or addition of high thermal con-
absorption rate is faster, the heat release is also more important ductivity materials). The challenge was to obtain fast apparent ab-
which leads to a higher peak temperature. sorption and desorption rates, through better internal and external heat
transfer management, while maintaining a large hydrogen volumetric
Regarding the influence of the initial and cooling fluid temperature capacity. The first configuration composed of a loose Ti-Mn powder
(Tinitial and Tf), the simulation results show that: (configuration 1) was considered as a Reference configuration.
Afterwards, compacts of Ti-Mn alloy powder (configuration 2:
- The hydrogen storage capacity (Fig. 8) is larger when the tem- Compact), an alternation of Ti-Mn alloy compacts and stainless steel
perature decreases (again in agreement with the absorption iso- fins (configuration 3: Alternation) and compacts of a mixture of Ti-Mn
therms described by Eqs. (21) and (22) in the study of Herbrig et al. alloy and stainless steel powders (configuration 4: Compact mixture)
[36]). This storage capacity is almost constant between 35 and were compared to evaluate their respective influence on the overall
55 ∘C, and clearly decreases above 55 ∘C; tank performance. The comparison was based on the absorption rate
- The hydrogen absorption rate (Fig. 7.d) is faster at low tempera- (apparent kinetics) as well as the total hydrogen storage capacity and
tures. As explained in Sections 4.1 and 4.2 of this paper, the bed ease of implementation.
temperature increases rapidly at the beginning of the absorption as a External heat transfer was studied by varying the value of the ex-
result of the exothermicity of the reaction. As a consequence, the ternal heat transfer coefficient (h) on the Reference configuration. For
apparent hydrogenation kinetics can be limited by the reaction the studied system, the implementation of an external heat transfer
equilibrium and the internal and external heat transfers need to be improvement technique (such as the addition of a double wall with
improved to obtain faster apparent kinetics. Decreasing the cooling circulation of calorific fluid, or addition of external fins) allowed rapid
fluid temperature increases the temperature gradient (between the cooling (heating) during absorption (desorption). However, the ob-
tank walls and the cooling fluid), which improves the external heat tained results showed that the external heat transfer had a negligible
transfer and leads to faster absorption rates; impact on the peak temperature observed at the beginning of the hy-
- The peak temperature observed at the beginning of the hydrogen drogen absorption or desorption process. Indeed, the low thermal
absorption is larger at high temperatures (Fig. 7.c). Indeed, starting conductivity of the metal hydride bed did not allow an efficient heat
from a higher initial temperature and using a higher cooling fluid supply / dissipation. Thus, an analysis on the improvement of the in-
temperature leads to higher temperatures during the hydrogenation ternal heat transfer had been performed. Two techniques were studied
reaction and also a larger peak temperature. and compared:

10
M.S. Manai, et al. Journal of Energy Storage 26 (2019) 101024

- The compaction of the metal hydride powder; Int. J. Hydrogen Energy 35 (2010) 912–919.
- The addition of a high thermal conductivity material in the form of [17] M.V. Lototskyy, I. Tolj, L. Pickering, C. Sita, F. Barbir, V. Yartys, The use of metal
hydrides in fuel cell applications, Prog. Nat. Sci. Mater. Int. 27 (2017) 3–20.
fins or mixed powder. [18] Z. Liu, Y. Li, Q. Bu, C.J. Guzy, Q. Li, W. Chen, C. Wang, Novel fuel cell stack with
coupled metal hydride containers, J. Power Sources 328 (2016) 329–335.
The simulation results showed that the increase of the metal hydride [19] J. Bellosta von Colbe, J.-.R. Ares, J. Barale, M. Baricco, et al., Application of hy-
drides in hydrogen storage and compression: achievements, outlook and perspec-
bed thermal conductivity promoted the internal heat transfer char- tives, Int. J. Hydrogen Energy 44 (2019) 7780–7808.
acteristics, which improved the overall hydrogen absorption rate. The [20] M. Bielmann, U.F. Vogt, M. Zimmermann, A. Zuttel, Seasonal energy storage system
addition of a second material with high thermal conductivity was based on hydrogen for self sufficient living, J Power Sources 196 (2011)
4054–4060.
shown to be an efficient heat transfer enhancement technique. The [21] C. Song, L.E. Klebanoff, T.A. Johnson, B.S. Chao, A.F. Socha, J.M. Oros, et al., Using
metal hydride storage configurations (Alternation and Compact metal hydride H2 storage in mobile fuel cell equipment: design and predicted
mixture) composed by both materials (metal hydride and stainless performance of a metal hydride fuel cell mobile light, Int. J. Hydrogen. Energy 39
(2014) 14896–14911.
steel), showed two advantages: the improvement of internal heat
[22] L. Valverde, F. Rosa, C. Bordons, J. Guerra, Energy management strategies in hy-
transfer and the increase of the hydrogen storage capacity (for a given drogen smart-grids: a laboratory experience, Int. J. Hydrogen. Energy 41 (2016)
tank volume). 13715–13725.
Compared to the storage bed composed by an alternation of metal [23] W. Liu, C.J. Webb, E.M. Gray, Review of hydrogen storage in AB3 alloys targeting
stationary fuel cell applications, Int. J. Hydrogen. Energy 41 (2016) 3485–3507.
hydride compacts and stainless steel fins (configuration 3: [24] T. Lohner, A. D'Aveni, Z. Dehouche, P. Johnson, Integration of large-scale hydrogen
Alternation), the compacted powder mixture (configuration 4: storages in a low-carbon electricity generation system, Int. J. Hydrogen Energy 38
Compact mixture) would be more difficult to separate and recycle. (2013) 14638–14653.
[25] A.T. Thattai, B.J. Wittebrood, T. Woudstra, J.J.C. Geerlings, P.V. Aravind,
Therefore, the concept of alternating compacted storage material Thermodynamic system studies for a natural gas combined cycle (NGCC) plant with
powders with high thermal conductivity fins (Alternation configura- CO2 capture and hydrogen storage with metal hydrides, Energy Procedia 63 (2014)
tion) gave the most promising results and offered the best compromise 1996–2007.
[26] G.L. Soloveichik, Regenerative fuel cells for energy storage, Proc. IEEE 102 (2014)
between the total hydrogen storage capacity, absorption rate and ease 964–975.
of implementation. [27] T. Yoshida, K. Kojima, Toyota MIRAI fuel cell vehicle and progress toward a future
Finally, various simulations were performed with the Alternation hydrogen society, Electrochem. Soc. Interfac. 24 (2015) 45–49.
[28] S. Metz, Linde pioneers hydrogen compression techniques for fuel cell electric ve-
configuration in order to investigate the influence of the hydrogen hicles, Fuel Cells Bull 2014 (2014) 12–15.
pressure and the cooling fluid temperature. A high pressure associated [29] X. Wang, R. Chen, Y. Zhang, C. Chen, Q. Wang, Hydrogen storage alloys for high-
to a low cooling fluid temperature gave the fastest hydrogen loading pressure suprapure hydrogen compressor, J. Alloy Comp 420 (2006) 322–325.
[30] M. Afzal, R. Mane, P. Sharma, Heat transfer techniques in metal hydride hydrogen
rates. From a practical point of view, the reference operating conditions
storage: a review, Int. J. Hydrogen Energy 42 (2017) 30661–30682.
(40 bar and 50 ∘C) chosen in this work were reasonable as they allowed [31] M. Afzal, P. Sharma, Design of a large-scale metal hydride based hydrogen storage
a high hydrogen capacity. reactor: simulation and heat transfer optimization, Int. J. Hydrogen Energy (2018)
1–17.
[32] S.S. Bhogilla, Design of a AB2-metal hydride cylindrical tank for renewable energy
Declaration of Competing Interest storage, J. Energy Storage 14 (2017) 203–210.
[33] C.N. Ranong, M. Höhne, J. Franzen, J. Hapke, G. Fieg, M. Dornheim, N. Eigen,
The authors declare that they have no known competing financial J.M. Bellosta von Colbe, O. Metz, Concept, design and manufacture of a prototype
hydrogen storage tank based on sodium alanate, Chem. Eng. Technol. 32 (2009)
interests or personal relationships that could have appeared to influ- 1154–1163.
ence the work reported in this paper. [34] S. Mellouli, F. Askri, H. Dhaou, A. Jemni, S. Ben Nasrallah, Numerical study of heat
exchanger effects on charge/discharge times of metal-hydrogen storage vessel, Int.
J. Hydrogen Energy. 34 (2009) 3005–3017.
References [35] H. Dhaou, A. Souahlia, S. Mellouli, F. Askri, A. Jemni, S. Ben Nasrallah,
Experimental study of a metal hydride vessel based on a finned spiral heat ex-
[1] N. Nemeth, Environment and energy : problems, resolutions, solutions, Int. J. changer, Int. J. Hydrogen Energy 35 (2010) 1674–1680.
Hydrogen Energy 15 (1990) 457–462. [36] K. Herbrig, L. Röntzsch, C. Pohlmann, T. Weißgärber, B. Kieback, Hydrogen storage
[2] P. Selvam, Energy and environment-An all time search, Int. J. Hydrogen Energy. 16 systems based on hydride-graphite composites: computer simulation and experi-
(1991) 35–45. mental validation, Int. J. Hydrogen Energy 38 (2013) 7026–7036.
[3] M. Paskevicius, D.A. Sheppard, K. Williamson, C.E. Buckley, Metal hydride thermal [37] M. Dietrich, C. Pohlmann, I. Burger, M. Linder, L. Rontzch, Long-term cycle stability
heat storage prototype for concentrating solar thermal power, Energy 88 (2015) of metal hydride-graphite composites, Int. J. Hydrogen Energy 46 (2015)
469–477. 16375–16382.
[4] G. Luderer, R.C. Pietzcker, S. Carrara, H.S. de Boer, S. Fujimori, N. Johnson, [38] I. Burger, M. Dietrich, C. Pohlmann, L. Rontzch, M. Linder, Standardized hydrogen
S. Mima, D. Arent, Assessment of wind and solar power in global low-carbon energy storage module with high utilization factor based on metal hydride-graphite com-
scenarios: an introduction, Energy Econ. 64 (2017) 542–551. posites, J. Power sources 342 (2017) 970–979.
[5] T.A. Semelsberger, K.P. Brooks, Chemical hydrogen storage material property [39] K. Herbrig, C. Pohlmann, L. Gondek, H. Figiel, N. Kardjilow, et al., Investigations of
guidelines for automotive applications, J. Power Sources 279 (2015) 593–609. the structural stability of metal hydride composites by in-situ neutron imaging, J.
[6] C.L. Aardahl, S.D. Rassat, Overview of systems considerations for on-board che- power sources 293 (2015) 109–118.
mical hydrogen storage, Int. J. Hydrogen Energy 34 (2009) 6676–6683. [40] K. Manickmanm, D.M. Grant, G.S. Walker, Optimization of AB2 type alloy com-
[7] A.V. Safronov, S.S. Jalisatgi, H.B. Lee, M.F. Hawthorne, Chemical hydrogen storage position with superior hydrogen storage properties for stationary applications, Int.
using polynuclear borane anion salts, Int. J. Hydrogen Energy. 36 (2011) 234–239. J. Hydrogen Energy 40 (2015) 16288–16296.
[8] A. Zuttel, A. Borgshulte, L. Schlapbach, Hydrogen As a Future Energy Carrier, [41] V.M. Skripnyuk, M. Ron, Evaluation of kinetics by utilizing the normalized pressure
Wiley-VCH, 2008. dependence method for the alloy Ti0.95Zr0.05Mn1.48V0.43Fe0.08Al0.01, J. of
[9] M. Wietschel, M. Ball, The future of hydrogen – opportunities and challenges, Int, J. Alloys and Coumpounds. 295 (1999) 385–390.
Hydrogen Energy 34 (2009) 615–627. [42] A. Jemni, S. Nasrallah, J. Lamloumi, Experimental and theoretical study of a metal-
[10] J. Zheng, X. Liu, P. Xu, P. Liu, Y. Zhao, J. Yang, Development of high pressure hydrogen reactor, Int. J. Hydrogen Energy 24 (1999) 631–644.
gaseous hydrogen storage technologies, Int. J. Hydrogen Energy 37 (2012) [43] G. Gwak, S. Yun, H. Ju, Multi-dimensional modeling and large-scale simulation of
1048–1057. hydrogen absorption/desorption phenomena in metal hydride vessels (MHVs),
[11] G. Valenti, Hydrogen liquefaction and liquid hydrogen storage, Compendium Fusion Eng. Des. 130 (2018) 107–113.
Hydrogen Energy 23 (2016) 28–51. [44] F. Askri, M. Ben Salah, A. Jemni, S. Ben Nasrallah, Optimization of hydrogen sto-
[12] R. Bhattacharyya, S. Mohan, Solid state storage of hydrogen and its isotopes: an rage in metal-hydride tanks, Int. J. Hydrogen Energy 34 (2009) 897–905.
engineering overview, Renew. Sustain. Energy Rev. 41 (2015) 872–883. [45] N.B. Selvaraj, D. Chapelle, D. Perreux, H. Figiel, Modelling the evolution of tem-
[13] B. Sakintuna, F. Lamari-Darkrim, M. Hirscher, Metal hydride materials for solid perature inside LaNi4.78Sn0.22 storage tank during refueling, Mater. Des. 30
hydrogen storage: a review, Int. J. Hydrogen Energy 32 (2007) 1121–1140. (2009) 954–957.
[14] F.E. Lynch, Metal hydride practical applications, J. Less Common Mater. 172 (1991) [46] M. Valizadeh, M.A. Delavar, M. Farhadi, Numerical simulation of heat and mass
943–958. transfer during hydrogen desorption in metal hydride storage tank by Lattice
[15] K. Sapru, S. Ramachandran, An integrated PV – electrolysis metal hydride hydrogen Boltzmann method, Int. J. Hydrogen Energy 41 (2015) 1–12.
generation and storage system, Energy Convers (2000). [47] P. Stephan, S. Kabelac, M. Kind, H. Martin, D. Mewes, K. Schaber, VDI Heat Atlas,
[16] V. Pérez-Herranz, M. Pérez-Page, R. Beneito, Monitoring and control of a hydrogen second ed., VDI e.V., Germany, 2010, p. 15.
production and storage system consisting of water electrolysis and metal hydrides,

11