121 (2024) 205196

Contents lists available at ScienceDirect

Gas Science and Engineering

journal homepage: www.journals.elsevier.com/gas-science-and-engineering

Modelling underground hydrogen storage: A state-of-the-art review of

fundamental approaches and findings
Motaz Saeed a, b, Prashant Jadhawar a, *
a
School of Engineering, University of Aberdeen, Aberdeen, AB24 3UE, Scotland, UK
b
Craft & Hawkins Department of Petroleum Engineering, Louisiana State University, Baton Rouge, 70803, USA

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

Keywords: This review presents a state-of-the-art of geochemical, geomechanical, and hydrodynamic modelling studies in
Energy storage the Underground Hydrogen Storage (UHS) domain. Geochemical modelling assessed the reactivity of hydrogen
Energy transition and res pective fluctuations in hydrogen losses using kinetic reaction rates, rock mineralogy, brine salinity, and
Geochemical modelling
the integration of hydrogen redox reactions. Existing geomechanics studies offer an array of coupled hydro-
Geomechanical modelling
Machine learning
mechanical models, suggesting a decline in rock failure during the withdrawal phase in aquifers compared to
Numerical modelling and simulation injection phase. Hydrodynamic modelling evaluations indicate the critical importance of relative permeability
Renewable energy hysteresis in determining the UHS performance. Solubility and diffusion of hydrogen gas appear to have minimal
impact on UHS. Injection and production rates, cushion gas deployment, and reservoir heterogeneity however
significantly affect the UHS performance, stressing the need for thorough modelling and experimental studies.
Most of the current UHS modelling efforts focus on assessing the hydrodynamic aspects which are crucial for
understanding the viability and safety of UHS. In contrast, the lesser-explored geochemical and geomechanical
considerations point to potential research gaps. A variety of modelling software tools such as CMG, Eclipse,
COMSOL, and PHREEQC evaluated those UHS underlying effects, along with a few recent applications of data-
driven-based Machine Learning (ML) techniques for enhanced accuracy.
This review identified several unresolved challenges in UHS modelling: pronounced lack of expansive datasets,
leading to a gap between model predictions and their practical reliability; need robust methodologies capable of
capturing natural subsurface heterogeneity while upscaling from precise laboratory data to field-scale conditions;
demanding intensive computational resources and novel strategies to enhance simulation efficiency; and a gap in
addressing geological uncertainties in subsurface environments, suggesting that methodologies from oil reservoir
simulations could be adapted for UHS.
This comprehensive review offers a critical synthesis of the prevailing approaches, challenges, and research
gaps in the domain of UHS, thus providing a valuable reference document for further modelling efforts, facili­
tating the informed advancements in this critical domain towards the realization of sustainable energy solutions.

1. Background transportation, and power generation (International Energy Agency,

The transition from fossil fuels to renewable forms of energy is of
paramount importance in mitigating possible climate change effects.
Burning fossil fuels releases large amounts of carbon dioxide and other
greenhouse gases into the atmosphere. Therefore, the development of
clean and sustainable energy sources has become a top priority for many
countries around the world. According to the International Energy
Agency (IEA), hydrogen can play a significant role in achieving global
climate goals by reducing carbon dioxide emissions from industry,
2022). Hydrogen can also help to increase energy security and reduce
the dependence on fossil fuels, particularly in countries that rely heavily
on imported oil and gas.
Hydrogen gas as an energy carrier has recently been integrated and
used in various applications, including transportation, power genera­
tion, and industrial processes (Abe et al., 2019; McCay and Shafiee,
2020). In the transportation sector, hydrogen fuel cells are being
developed as an alternative to internal combustion engines. Hydrogen
fuel cells are an attractive alternative to traditional fossil fuel-powered

* Corresponding author.
E-mail addresses: motazsaeed@lsu.edu (M. Saeed), Prashant.Jadhawar@abdn.ac.uk (P. Jadhawar).

https://doi.org/10.1016/j.jgsce.2023.205196
Received 31 August 2023; Received in revised form 9 December 2023; Accepted 12 December 2023
Available online 16 December 2023
2949-9089/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
M. Saeed and P. Jadhawar
vehicles due to their zero-emission property. Fuel cell vehicles emit only
water vapour and heat, making them an ideal solution for reducing
greenhouse gas emissions in the transportation sector. In addition,
hydrogen fuel cells offer longer driving ranges and shorter refuelling
times than battery electric vehicles (Abe et al., 2019; Ahluwalia et al.,
2005; Singla et al., 2021; Aminudin et al., 2023a).
Hydrogen energy is also gaining popularity in industrial applica­
tions, particularly in the steel and chemical industries. Hydrogen can be
used as a reducing agent in steel production, replacing coal and other
fossil fuels that are currently used (Liu et al., 2021). The global steel
sector is expected to potentially reduce its carbon emissions by up to
50% by replacing coal with hydrogen by 2050 (International Energy
Agency, 2022). Hydrogen is also an important feedstock in the pro­
duction of chemicals such as ammonia, methanol, and other chemicals.
The use of hydrogen in these industries has the potential to significantly
reduce greenhouse gas emissions.
One key aspect of hydrogen is its potential to act as an energy store
for excess energy generated from other renewable sources (Gallo et al.,
2016). For instance, during periods of low energy demand, excess en­
ergy produced from wind and solar power can be converted into
hydrogen gas and stored for later use during peak energy demand pe­
riods. This can help to stabilize the energy supply and ensure that energy
is available whenever needed mitigating renewable energy supply
fluctuations (Gallo et al., 2016; Tarkowski, 2019).
Hydrogen can be produced from various sources, including fossil
fuels such as natural gas and coal, as well as renewable sources such as
wind, solar and water (Aminudin et al., 2023b). Based on the production
process, produced hydrogen can be classified as grey hydrogen, brown
hydrogen, blue hydrogen, or green hydrogen as depicted in Fig. 1. Grey
hydrogen is produced through the gasification of coal which produces
CO2 as a by-product and is emitted to the atmosphere (Hermesmann and
Müller, 2022). The most common method of producing hydrogen is
through steam methane reforming (SMR), which involves the reaction of
methane with steam in the presence of a catalyst. This process is
currently the most widely used method of hydrogen production and
accounts for approximately 95% of the world’s hydrogen production
(International Energy Agency, 2022). However, the process of SMR
produces greenhouse gas emissions, which makes it less environmen­
tally friendly. Thus, technologies such as carbon capture, utilization, and
storage (CCUS) have been developed to capture and store the carbon
dioxide emissions from SMR. The hydrogen gas produced from the SMR
process is either known as brown hydrogen when the produced CO2 is
emitted to the atmosphere or blue hydrogen when the CO2 is captured
(Hermesmann and Müller, 2022; Dawood et al., 2020).
Another method of hydrogen production is through electrolysis,
which involves splitting water molecules into hydrogen and oxygen
Fig. 1. Chart showing the types of hydrogen, their production sources, processes and respective CO2 mitigation options.
2
Gas Science and Engineering 121 (2024) 205196
using electricity. Electrolysis can use renewable energy sources such as
wind and solar power to produce hydrogen without greenhouse gas
emissions (Yu et al., 2021). This process produces what is known as
green hydrogen and has the potential to become an essential component
of the clean energy transition. It is estimated that low-emission
hydrogen production could reach 24 million tonnes by 2030, up from
the current production of less than 1 million tonnes (International En­
ergy Agency, 2022).
One important aspect of the hydrogen economy and supply chain is
hydrogen storage (Gallo et al., 2016). UHS in geological formations is a
promising method for storing hydrogen gas during periods of low energy
demand and releasing it when energy demand increases (see Fig. 2).
Compared to above-ground storage options, UHS offers several advan­
tages, including the ability to store hydrogen in the large quantities
necessary for widespread use as an energy source, increased safety due
to reduced likelihood of gas leaks, and the potential for high-pressure
storage to increase energy density and quality enhancing its suitability
for use in fuel-cell cars (Zivar et al., 2021a; Flesch et al., 2018).
Aquifers offer several attractive features for hydrogen storage, such
as abundance, stability, and the ability to store hydrogen at high pres­
sures. When hydrogen is injected into an aquifer, it dissolves into the
water to some extent, is subsequently stored in the rock’s pores and
cracks, then ready for later retrieval to satisfy the energy needs. On the
other hand, salt caverns as an underground storage option offer several
benefits due to their tightness and low reactivity with the hydrogen, thus
reducing the possibility of hydrogen loss due to geochemical reactions.
Additionally, their visco-plastic properties make them well-suited for
sealing gas movement and avoiding mechanical failures. However, the
construction of salt caverns requires a source of fresh water for solution
mining (Muhammed et al., 2022; Laban, 2020).
Depleted oil and gas reservoirs also represent viable promising op­
tions for UHS due to the industrial operators’ experience in injecting
water and CO2, hence the information on reservoir rock and fluid
properties, pressure regime, storage capacity, connectivity and geo­
mechanical effects on the storage system (Hematpur et al., 2023; Lew­
andowska-Śmierzchalska et al., 2018). Possible reactions between
hydrogen and residual fluids may affect the purity of the stored
hydrogen (Hemme and van Berk, 2018). Therefore, it is necessary to
carefully evaluate each storage option to ensure the safety and reliability
of hydrogen storage. Nonetheless, UHS remains an attractive option for
the transition to a hydrogen-based economy to achieve the ongoing
research objectives to further improve and expand the technology.
The present knowledge and understanding of UHS have been derived
from underground gas storage of other gases such as carbon dioxide and
natural gas (Zivar et al., 2021a; Muhammed et al., 2022). However,
hydrogen gas has different properties that require an alteration in the
M. Saeed and P. Jadhawar
Fig. 2. Schematic representation of UHS.
underground gas storage process to prevent metal embrittlement from
H2. Current operating UHS projects are primarily located in salt caverns,
however, there is a shift towards aquifers and depleted hydrocarbon
reservoirs due to their high capacity and available geological informa­
tion and models, especially in the case of depleted hydrocarbon reser­
voirs. More details on the operating and pilot studies of UHS can be
found in other review papers that have overall focused on underground
hydrogen storage (Zivar et al., 2021a; Hematpur et al., 2023; Muham­
med et al., 2023; Sambo et al., 2022).
Due to the growing importance of cleaner energy sources, UHS is
becoming increasingly significant. However, as most of the current
projects are still in the research or pilot stage, it will be a while before
this technology can be implemented on a large scale. To expedite the
development of UHS, modelling and simulation studies play a crucial
role. These studies offer a cost-effective way to evaluate the feasibility of
UHS and optimise the design of storage facilities, reducing the risk of
expensive errors during the construction and operation of UHS projects.
In addition, these studies can identify areas that require further research
to improve the safety and performance of UHS.
This publication focuses specifically on critically reviewing and
analysing the existing modelling studies of UHS with the objective of
providing a clear and concise understanding of the current state of UHS
modelling. The review systematically analyses and categorizes the
research into three main categories: geochemical, geomechanical, and
hydrodynamic aspects. Within each category, the paper investigates the
fundamental concepts, process involving equations, modelling ap­
proaches, and findings providing a thorough and accurate overview of
the existing literature. Additionally, the review dedicates a section to
exploring emerging trends in UHS modelling. By methodically
comparing various modelling approaches and studies, this paper effec­
tively highlights the current status of modelling efforts and underscores
the gaps and challenges faced in the UHS field. The paper concludes by
drawing a well-informed conclusion and suggesting the areas for
improvement in future modelling efforts.
This information will prove to be valuable for researchers, engineers,
and policymakers who are working on the development and imple­
mentation of UHS. By building on the knowledge gained from previous
modelling studies, we can accelerate the development of this important
UHS technology and help to pave the way for a more sustainable energy
future.
3
Gas Science and Engineering 121 (2024) 205196
2. UHS modelling aspects and approaches
UHS is a complex process that is affected by different mechanisms
operational during the geochemical, geomechanical and hydrodynami­
cal interactions in the formation. These aspects have a significant impact
on the hydrogen gas injection, storage, and production/withdrawal
stages of UHS. Moreover, these mechanisms and aspects interact syn­
ergistically during UHS making it more complicated to predict their
overall impact on the UHS. Therefore, it is essential to understand the
individual and collective effects of these mechanisms to properly predict
the performance of UHS. Furthermore, the properties and parameters
associated with of the studied aspect, geochemical, geomechanical and
hydrodynamical, require careful calculation, measurement, or estima­
tion. These properties and parameters can be rock properties, fluid
properties, operational parameters, geomechanical properties, reservoir
conditions, microbial conditions, among others. A summary of the most
important parameters, properties and mechanisms operational during
UHS is shown in Fig. 3 and further details are discussed in the next
sections. The Various modelling approaches have been proposed and
undertaken to investigate these effects. A detailed description of the
effects and the undertaken modelling approaches in the previous studies
will be presented in the following subsections.
2.1. Geochemical
In this section the geochemical aspects of UHS with respect to fun­
damentals, modelling approaches, and previous studies findings will be
discussed. Various parameters should be considered in geochemical
modelling these include pressure and temperature of the reservoir, the
resident fluid’s composition, salinity and pH, and the type of microor­
ganisms and their rate of growth. These parameters affect various
geochemical phenomena that define the overall geochemical impact
during UHS, and these are abiotic reactions, biotic reactions, and min­
eral dissolution and precipitation reactions. A summary of the most
important parameters and phenomena for geochemical modelling is
depicted in Fig. 4.
2.1.1. Fundamentals and modelling approaches
Geochemical reactions during UHS may occur between the stored
hydrogen gas, cushion gas, reservoir fluids and reservoir minerals (Yekta
et al., 2018). A cushion gas is a gas that is injected into the storage zone
M. Saeed and P. Jadhawar Gas Science and Engineering 121 (2024) 205196

Fig. 3. Aspects of modelling UHS. The dark blue area represents the most important aspects in modelling UHS. The light yellow area represents the main mechanisms
and phenomena associated with each aspect. The light blue area details the most important parameters and properties in modelling UHS.

Fig. 4. Most important parameters and phenomena in modelling the geochemical aspects of UHS.

before the injection of the hydrogen gas for the purposes of providing the stored hydrogen due to converting hydrogen to other gases such as
pressure support and limiting the spread of hydrogen gas inside the methane and hydride sulfide. Moreover, geochemical reactions have the
reservoir. Investigated cushion gases involve CH4, N2, and CO2 (Saeed potential of reducing the purity of hydrogen gas at the time of produc­
and Jadhawar, 2023). The geochemical reactions may lead to losses in tion due to mixing with other gases present in the reservoir. The

4
M. Saeed and P. Jadhawar Gas Science and Engineering 121 (2024) 205196

hydrogen gas purity refers to percentage of hydrogen gas produced Table 1

during the withdrawal stage. This purity can be affected by the pro­ A selection of possible geochemical reactions during UHS (Hemme and van Berk,
duction of associated gases such as the utilised cushion gas, residual gas 2018; Korrani et al., 2016).
in the case of storage in depleted gas reservoirs or gases produced as a Reaction Log K Reaction Log K oC25
result of the geochemical reactions involving hydrogen gas which may o
25
C

happen over a prolonged period of time depending on the storage con­ Equilibrium Reactions Ion Exchange Reactions
ditions (Muhammed et al., 2023). Another important possible effect of
H2 O ↔ H + OH
+ − − 14.0 Na+ + X− ↔ NaX 0
the geochemical activity of hydrogen gas during UHS is the possibility of − 22.7 0.8
Al3+ + 4H2 O ↔ Ca2+ + 2X− ↔ CaX2
the geological structure’s mineral dissolution or precipitation leading to Al(OH)−4 + 4H+
a change in its porosity and permeability. Consequently, the geological Al3+ + 3H2 O ↔ − 16.9 Mineral Dissolution/Precipitation Reactions
structure’s geomechanical properties may be altered resulting in Al(OH)3 + 3H+
possible structure failure or faulting and reducing cap rock integrity Al3+ + 2H2 O ↔ − 10.1 KAlSi3 O8 + 8H2 O ↔ K+ + − 20.573
(Hemme and van Berk, 2018). Geochemical reactions (see Fig. 5 and Al(OH)+ Al(OH)−4 + 3H4 SiO4
2 + 2H
+

Table 1) during UHS can either be abiotic or biotic reactions. Abiotic CO2−
3 + H ↔ HCO3
+ − − 5.0 NaAlSi3 O8 + 8H2 O ↔ Na+ + − 18
Al(OH)−4 + 3H4 SiO4
reactions are those that occur between hydrogen and non-living or­
CO2−
3 + 2H ↔
+ 10.329 Al2 Si2 O5 (OH)4 + 6H+ ↔ H2 O + 7.435
ganisms such as brine, oil, gas, and rock minerals. Abiotic reactions can
CO2 + H2 O 2H4 SiO4 + 2Al3+
be equilibrium reactions, mineral dissolution/precipitation reactions, or CO2− 16.681 SiO2 + 2H2 O ↔ H4 SiO4 3.95
3 + Ca
2+
+
ion exchange reactions (Zivar et al., 2021a). Equilibrium reactions are H+ ↔ CaHCO+ 3
fast reactions that are both forward and reverse reactions such that the HCO−3 + 11.935 CaCO3 ↔ CO2−
3 + Ca
2+ 8.48
concentrations of the reactants and products remain constant over time. Na+ ↔ NaHCO3
They can be written in the general form of Eq. (1). CO2−
3 + Mg
2+
+ − 0.25 FeS2 + 2H+ + 2e− ↔ Fe2+ + 2HS− − 18.479
H+ ↔ MgHCO+ 3
aR + bS→ cP 1 CO2−
3 +
11.399 K0.6 Mg0.25 Al2.3 Si3.5 O10 (OH)2 + − 40.267
Na+ ↔ NaCO−3 11.2H2 O ↔ 0.6K+ 0.25Mg2+ +
where a, b and c are the stoichiometric coefficient of geochemical spe­ 2.3Al(OH)4− + 3.5H4 SiO4 + 1.2 H+
cies R, S and P, respectively. Their equilibrium constants (K) can be CO2−
3 + 1.27 CaMg(CO3 )2 ↔ Ca2+ + Mg2+ + − 17.09
calculated using Eq. 2. Ca2+ ↔ CaCO3 2CO2−
3
CO2−
3 +
3.224 Biotic Reactions
acTP Mg2+ ↔ MgCO3
K= 2
aaR .abS Ca2+ + H2 O ↔ 2.98 CO2 + 4H2 →CH4 + 2H2 O − 31
CaOH+ + H+
where a is the activity of their respective chemical species. The activities Mg2+ + H2 O ↔ − 12.78 2CO2 + 4H2 →CH3 COOH + 2H2 O − 12
are related to the molar concentrations of their species through their MgOH+ + H+
Na+ + OH+ ↔ NaOH − 11.44 SO2−
4 + 5H2 →H2 S + 4H2 O
35
individual activity coefficients as shown in Eq. (3).
ai = γi .Ci 3
WATEEQ Debye-Huckel model (Eq. (4)) can be used to calculate the and bi are ion specific parameters fitted from mean-salt activity coeffi­
activity coefficients: cient (Parkhurst and Appelo, 2013a). A and B are constant their values
√̅̅ depend on temperature. The ionic strength is a function of the concen­
log γ i = −
Az2i I
√̅̅ − bi I 4 tration of all the ions present in the solution and is expressed as in Eq.
1+ I (5).

where I is the ionic strength and for WATEQ Debye-Huckel equation, Ci

Fig. 5. An example of geochemical activity during UHS (after Hemme et al. (Hemme and van Berk, 2018)).

5
M. Saeed and P. Jadhawar
∑
n Sulfate reduction reactions convert H2 and sulfate to hydrogen sul­
I = 0.5 ci z2i
i=1
where ci is the molar concentration of ion i and z is the valence number
of that ion i. The hydrogen solubility in water can be written as a
geochemical reaction in the form of Eq. (6).
H2(g) ↔ H2(aq)
where the subscripts (g) and (aq) denote that the hydrogen is present in
gaseous phase or aqueous phase, respectively. The equilibrium constant
for this reaction can be considered Henery’s constant which relates the
partial pressure of the gas with the concentration of hydrogen dissolved
in the solution and depends on both the pressure and temperature
conditions. The other equilibrium reactions that occur in a H2 geologic
storage system (examples shown Table 1), even if they do not involve
hydrogen directly, affect the pH and chemistry of the solution which
directly affects the solubility of H2 in water. Increasing the ionic
strength of a solution, such as by increasing its salinity, often leads to a
"salting-out" effect for many gases. This effect results in decreased gas
solubility in the solution.
Ion exchange reactions (examples in Table 1) are also reversible re­
actions where a free ion in a solution replaces another ion that is bound
to a solid surface or solution. The activity coefficient ion exchange re­
actions can be calculated using WATEEQ Debye-Huckel model. The use
of selectivity coefficients follows the Gaines-Thomas convention (Gaines
and Thomas, 1953). The cation exchange capacity (CEC) that quantifies
the ion exchange is defined as the measure of how many cations can be
retained on soil particle surfaces (Brady and Weil, 2016) and is related to
the equivalent fractions using Eq. (7):
zi ci zi ci
ζi = N = 7
∑ ex CEC
zi ci
i=1
where ci is the concentration of the exchange species, zi is the valence
number of the exchange species, and Nex is the total number of exchange
species.
Mineral dissolution and precipitation reactions are the kinetically
controlled heterogeneous solubility product reactions. The rate of min­
eral dissolution and precipitation reaction per unit bulk volume of
porous medium (mol/(m3⋅s)), rβ is expressed in Eq. (8) (Bethke, 1996)
( )
̂ β kβ 1− Qβ , β = 1…, n (r, m) n
rβ = A 8 complexation reactions, 1D and 2D reactive and colloidal transport
Keq,β
where Âβ is the reactive surface area of reactant mineral β per unit bulk
volume of porous medium (m2/m3), kβ is the rate constant of mineral
reaction β (mol/m2⋅s). Kβ is the chemical equilibrium constant of min­
eral reaction, Qβ is the activity product of reaction and nr,mn is the
number of mineral reactions.
Hydrogen biotic reactions as a result of microbial activities are one of
the key aspects that needs to be considered during UHS. These reactions
are dictated by the presence and type of living microorganisms (e.g.
bacteria) inside the storing site (Thaysen et al., 2021). Some of the know
biotic processes involving the consumption of hydrogen are methano­
genesis, acetogenesis, sulfate reduction and iron reduction. Methano­
genesis is the conversion of hydrogen to methane and water when
reacting with CO2 in the presence of methanogens (Panfilov, 2016). It is
also known as the Sabatier reaction and is shown in Eq. 9
CO2 + 4H2 → CH4 + 2H2 O
Acetogenesis is another type of biotic reactions that occurs between
H2 and CO2 in the presence of acetogen bacteria. This reaction produces
acetic acid and water as shown in Eq. (10) (Rabii et al., 2019; Gregory
et al., 2019)
2CO2 + 4H2 → CH3 COOH + 2H2 O 10
Gas Science and Engineering 121 (2024) 205196
5 fide (H2S) in the presence of sulfate-reduction bacteria (SRB) (Thaysen
et al., 2021; Cavallaro et al., 2005). SRB are mostly active in tempera­
tures around 38 ◦ C (Bernardez et al., 2013) but have been reported to be
active at temperatures up to 110 ◦ C (Machel, 2001). However, it is ex­
pected that its activity is slower in acidic aqueous and depleted hydro­
carbon reservoirs, especially in the presence of high salinity brine.
6
SO2−4 + 5H2 → H2 S + 4H2 O 11
Most of the existing studies addressing the modelling of the
geochemical and biochemical effects during UHS used US Geological
Survey’s PHREEQC. PHREEQC is a state-of-the-art geochemical model­
ling software that allows for the modelling of aqueous, mineral, ion-
exchange and surface complexation reactions (Parkhurst and Appelo,
2013a). To calculate solute activities, PHREEQC uses two
ion-association aqueous models (the Lawrence Livermore National
Laboratory model and WATEQ4F), Pitzer or, SIT (Specific ion Interac­
tion Theory) equations. Temperature effects on reactions equilibrium
constants are accounted for using Van’t Hoff equation (Eq. 12) (Sandler,
2006)
[( ) ( )]
ln K2 ΔH ◦ 1 1
= − 12
K1 R T1 T2
where K1 and K2 are the equilibrium constants at temperatures T1 and
T2 , respectively, ΔH◦ is the standard enthalpy change of the reaction, R
is the gas constant.
It is also possible to use analytical parameters to simulate tempera­
ture’s effect on equilibrium constants. PHREEQC enables batch-
reaction, 1D transport calculations and utilises the Peng-Robinson
equation of state (Eq. (13)) (Peng and Robinson, 1976) and Henry’s
Law to calculate gas solubilities at high pressures.
RT aα
13
P= =
Vm − b Vm2 − 2bVm − b2
where P is the pressure, T is the temperature, R is the gas constant, V is
the molar volume, a and b are empirical constants, and α is a
temperature-dependent parameter.
Geochemist’s workbench (GWB), a commercial software developed
by Aqueous Solutions LLC, that is also utilised to simulate geochemical
and biochemical reactions during UHS. GWB is capable of simulating
aqueous, kinetic mineral dissolution and precipitation and surface
(Bethke and FBSM, 2021). PHREEQC and GWB both are usually used for
geochemical modelling, however, GWB’s thermodynamics database is
constantly reviewed and updated as opposed to PHREEQC’s database
which uses publicly accessible thermodynamics databases. PHREEQC is
open source and hence its usage is free but may require some pro­
gramming knowledge to be able to write the geochemical codes.
2.1.2. Analysis of previous studies’ findings
Amid et al. (2016) modelled the storage of hydrogen in a depleted
natural gas reservoir considering the geochemical and biochemical ef­
fects. Their results showed that sulfate reduction pose a challenge to the
UHS process and is more impactful than methanation. However,
hydrogen loss due to dissolution was limited to less than 0.1%. The
modelled UHS case could supply 42% of the energy capacity supply
compared to the Rough Gas Storage Facility (UK) which is a natural gas
facility. This indicates that UHS can indeed provide a viable option for
9 energy storage. Hemme et al. (Hemme and van Berk, 2018) modelled the
geochemical and biochemical effects on the underground hydrogen
studies. They developed a model using PHREEQC to understand the
potential loss of hydrogen and changes in rock properties during the
storage process due to solubility, geochemistry, methanation, sulfate
reduction and diffusion. The authors included the effect of kinetics for
the hydrogen biotic reactions to understand their long-term effects on
6
M. Saeed and P. Jadhawar
the UHS. Results of their work showed that hydrogen loss due to diffu­
sion is minimal. While hydrogen loss as a result of biochemical reactions
is dependent on the methanation and sulfate reduction bacteria activity
at storage conditions. Another geochemical model was developed by
Hassannayebi et al. (2019) to investigate hydrogen geochemical re­
actions during UHS. They utilised the geochemical software Geo­
chemists Workbench and considered both equilibrium and mineral
reactions in a sandstone reservoir at 40 ◦ C and 7.5 bar. Their results
show that equilibrium reactions between hydrogen and aqueous-phase
components are more dominant during the storage-relevant time span.
Laban (2020) modelled the geochemical aspects of UHS in salt cav­
erns using PHREEQC. Laban’s work showed that bacteria growth rate,
sulfate concentration in brine, brine pH and ionic strength, cavern
pressure and temperature, and iron ions concentrations are the most
influencing factors on hydrogen sulfide production through sulfate
reduction bacteria. Bo et al. (2021) developed a geochemical model
using PHREEQC to evaluate the losses of hydrogen inside a sandstone
reservoir. They found that pressure and temperature play a negligible
role in hydrogen loss due to solubility. They also suggested that sand­
stone reservoirs would be better for hydrogen storage than carbonate
reservoirs because water uptake of hydrogen does not react with quartz
and clay minerals. Unlike carbonate reservoirs in which calcite dissolves
to trigger up to 9.5% hydrogen loss. Moreover, their modelling results
show that 87% of calcite would be dissolved over a 30-year period.
Therefore, they suggested that deep calcite-free reservoirs with
calcite-free caprocks would be better for UHS. Zeng et al. (2022a)
investigated the geochemical effects on UHS in Majiagou carbonate
reservoir using PHREEQC. They studied hydrogen loss, mineral disso­
lution and water chemistry over a long period of time. Their results
showed that hydrogen loss due to fluid-rock reactions up to 6.6% is
expected for the first year compared to 81.1% after 500 years of storage.
Calcite dissolution was negligible and limited to only 0.0646% over a
500-year period indicating that the Majiagou formation is a good
candidate for UHS. In another work, Zeng et al. (2022b) coupled
geochemical modelling using PHREEQC with disjoining pressure cal­
culations to evaluate the effects of various reservoir parameters on the
wettability of a hydrogen-brine-calcite system. Their modelling results
showed that increasing the temperature and decreasing the salinity in­
crease the disjoining pressure within the hydrogen-brine-calcite system
resulting in higher hydrogen gas repulsion from the calcite surface and
increasing water wettability. Tremosa et al. (2023) utilised PHREEQC to
evaluate the reactivity of hydrogen within UHS in the Lobodice town gas
storage site, focusing on its behaviour under various conditions,
including microbial influence. They reported that under abiotic condi­
tions, UHS shows minimal hydrogen reactivity. However,
microbial-driven processes like methanogenesis, sulfate-reduction, and
acetogenesis can lead to hydrogen consumption, yielding methane and
hydrogen sulfide. Moreover, methanogenesis, influenced by carbon di­
oxide and carbonates, significantly contributes to hydrogen losses.
Sulfate-reduction has a comparatively smaller effect on hydrogen con­
tent, generating hydrogen sulfide and associated corrosion risks. The
incorporation of microbial kinetics into models is essential for precise
simulations of hydrogen reactivity (Tremosa et al., 2023). The effect of
H2-brine-rock interactions on caprock integrity during UHS was inves­
tigated by Zeng et al. (2023) using kinetic batch models and analytical
estimates. Their findings revealed that the dissolution degrees of tested
minerals in different shale types are <1% over 30 years, indicating
robust caprock integrity. Their results affirmed that H2-brine-shale in­
teractions are unlikely to compromise caprock integrity during UHS.
The effects of geochemical reactions on hydrogen and reservoir rock
porosity and permeability during UHS in an aquifer have been investi­
gated by Saeed et al. (2023a). Their simulations through PHREEQC
revealed that geochemical reactions have a minimal impact on the
hydrogen gas in the simulated 30 years period. However, they observed
changes in the reservoir rock porosity and permeability as a result of the
feldspar and calcite dissolution especially when CO2 was used as a
7
Gas Science and Engineering 121 (2024) 205196
cushion gas.
The results of Amid et al. (2016), Hemme et al. (Hemme and van
Berk, 2018) and Saeed et al. (2023a) agree on their prediction of the
negligible hydrogen loss due to geochemical reactions and diffusion
during UHS which in in line with the experimental findings of Hassan­
pouryouzband et al. (2022) which indicated that there is no risk of
hydrogen abiotic reactivity in temperatures up to 80 ◦ C. Unlike the work
of Bo et al. (2021) and Zeng et al. (2022a) where they found that
hydrogen loss can be up to 9.5% and 81.1%, respectively, depending on
the storage duration, and reservoir mineral composition and conditions.
The variance in projected hydrogen depletion could potentially arise
from the inclusion of hydrogen redox reactions, which promote the
escalation of hydrogen solubility and subsequent loss. This phenomenon
aligns with the investigations conducted by Bo et al. (Saeed et al.,
2023a) and Zeng et al. (Hassanpouryouzband et al., 2022), in contra­
distinction to the findings of Amid et al. (Tremosa et al., 2023) and
Hemme et al. (Hemme and van Berk, 2018). The adoption of hydrogen
redox reactions and their rates is a matter of discussion and research at
the moment and requires further experimental work (Hassannayebi
et al., 2019). Zhan et al. (2023) investigated the redox reactions’ impact
on hydrogen-brine-mineral interactions in UHS. Minerals such as calcite,
siderite, quartz, and pyrite at varying dissolved oxygen levels (5.5–5500
ppm), temperature, and pressure were assessed by employing
geochemical modelling through PHREEQC. They found that dissolved
oxygen concentration amplification yields minimal influence on
hydrogen solubility and pH, while siderite and calcite exhibit reactivity,
incurring modest hydrogen loss under certain pressures. Similarly,
quartz and pyrite displayed limited reactivity. However, they found that
carbonate minerals provoke hydrogen dissociation and loss through
abiotic geochemical reactions, emphasizing sandstone reservoirs’ role in
mitigating hydrogen conversion and contamination concerns in UHS.
Calcite dissolution was predicted by Hemme et al. (Hemme and van
Berk, 2018), Bo et al. (2021), Zeng et al. (2022a), Zhan et al. (2023), and
Saeed et al. (2023a) as a result of the geochemical reactions between
hydrogen gas, cushion gas, calcite and brine. This implies that the ex­
istence of calcite within the geological formation could induce alter­
ations in the porosity and permeability of the rock, thereby influencing
the storage potential and hydraulic conductivity of the reservoir.
Moreover, the diminishment of hydrogen resulting from geochemical
reactions and diffusion exhibits a direct correlation with variations in
pressure and temperature. Elevated pressure and temperature levels are
conducive to heightened hydrogen losses during UHS operations
(Hemme and van Berk, 2018; Bo et al., 2021). A summary of the studies
on geochemical reactions effects during UHS is presented in Table 2.
Geochemical modelling of UHS is an essential aspect of modelling
the different aspects of UHS. In addition to the equilibrium reactions,
mineral dissolution/precipitation reactions should be properly accoun­
ted for in the model. The rates of kinetic reactions are dependent on
temperature, pressure and salinity (Hemme and van Berk, 2018). Hence,
the adoption of constant reaction rates in the modelling approaches
might lead to inaccurate results. Further, geochemical experiments
involving hydrogen reactions with various reservoir minerals, brine
composition and conditions are needed to calibrate and validate the
reaction rates and consequently geochemical modelling of UHS.
2.2. Geomechanical
In this section the geomechanical aspects of UHS will be discussed
with respect to fundamentals, modelling approaches, and the finding of
previous modelling studies. There are various properties and parameters
that impact the geomechanical behaviour of reservoir rock during UHS
such as the rock’s tensile, compression and shear stresses, presence of
faults and fractures, wellbore geometry, and injection and production
rates. As a result, various geomechanical phenomena and effects are
modelled including the evaluation of cap rock and wellbore integrity,
fracturing of reservoir formation, and faults and fractures reactivation. A
M. Saeed and P. Jadhawar Gas Science and Engineering 121 (2024) 205196

Table 2
A summary of studies on geochemical modelling of UHS.
Reference Scope Storage Approaches and Software Remarks
formation type packages

Bo et al. (2021) Geochemical modelling to evaluate the Sandstone PHREEQC Pressure and temperature play a negligible role in hydrogen
losses of hydrogen inside a sandstone reservoir loss due to solubility
reservoir Carbonate Sandstone reservoirs would be better for hydrogen storage
reservoir than carbonate reservoirs
Zeng et al. Evaluation of the effects of various Carbonate PHREEQC Increasing the temperature and decreasing the salinity
(2022b) reservoir parameters on the wettability reservoir DLVO Theory increase the disjoining pressure within the hydrogen-brine-
of a hydrogen-brine-calcite system calcite and increasing water wettability.
Amid et al. (2016) Investigate the geochemical and depleted natural – Sulfate reduction pose a challenge to the UHS process and is
biochemical effects on UHS gas reservoir more impactful than methanation hydrogen loss due to
dissolution was limited to less than 0.1%
Zeng et al. Investigate the geochemical effects on Majiagou PHREEQC Hydrogen loss due to fluid-rock reactions up to 6.6% is
(2022a) UHS in carbonate carbonate expected for the first year compared to 81.1% after 500 years
reservoir of storage
Laban (2020) Assess the geochemical aspects of UHS Salt cavern PHREEQC Bacteria growth rate, sulfate concentration in brine, brine
in salt caverns pH and ionic strength, cavern pressure and temperature, and
iron ions concentrations are the most influencing factors on
hydrogen sulfide production from hydrogen
Hemme and van Evaluation of geochemical reactions Depleted gas PHREEQC Hydrogen loss is mostly due to biotic reactions and are
Berk (2018) effects during UHS field dictated by the kinetic reaction rates
Hassannayebi Instigated effects of geochemical Sandstone Geochemical modelling using Equilibrium reactions between hydrogen and aqueous-phase
et al. (2019) reactions during UHS reservoir Geochemist Workbench components are more dominant during the storage-relevant
time span.
Ebigbo et al. Investigated the effects of bioactivity on N/A Mathematical and numerical Possible competition between three different microbial
(2013) UHS modelling species: methanogens, acetogens, and acetotrophs.
Hagemann et al. Investigate biochemical and Depleted gas Numerical modelling through The number of microorganisms present in the reservoir
(2016) hydrodynamic behaviour during UHS. reservoir DuMuX coupled with greatly influence the process of hydrogen conversion into
biochemical modelling methane.
Tremosa et al. Evaluate the geochemical impacts on Aquifer where Geochemical modelling using Microbial activity can have a significant effect on the stored
(2023) the stored hydrogen gas. town gas is PHREEQC hydrogen. Biotic reactions kinetics should be considered
stored carefully in modelling.
Zeng et al. (2023) Evaluate the effects of geochemical Depleted gas Geochemical modelling using Minimal shale minerals dissolution is expected during UHS
reactions on cap rock integrity during reservoir PHREEQC and hence integrity of seal is preserved against H2-brine-
UHS. rock geochemical reactions.
Saeed et al. Investigate the geochemical effects Deep aquifer Geochemical modelling using Minimal effect of geochemical reactions on the stored
(2023a) during UHS on storage gases and rock PHREEQC hydrogen gas. Rock porosity and permeability may increase
properties due to calcite and feldspars dissolution as a result of using
CO2 as a cushion gas.
Zhan et al. (2023) Studied the role of redox reactions in Aquifer Geochemical modelling using Carbonate minerals lead to more redox reactions and
UHS PHREEQC hydrogen disassociation. Carbonates are more reactive
during UHS compared to sandstone rocks. Hence, sandstone
rocks are preferred candidates.

summary of the primary parameters and phenomena for geomechanical

modelling is depicted in Fig. 6. ρ(∂v / ∂t + v ⋅ ∇v) = − ∇P + μ ∇2 v 14

where v is the fluid velocity, and the terms on the right-hand side
2.2.1. Fundamentals and modelling approaches
represent the pressure gradient and viscous stress, respectively (Blyton
During hydrogen storage, cycles of hydrogen injection and with­
et al., 2015). This conservation of momentum equation describes the
drawal are expected and with each cycle the reservoir pore pressure
fluid behaviour during H2 injection and production accounting for the
shifts leading to changes in the effective stress within the storage for­
pressure gradient, ∇P. To calculate this pressure gradient, Biot’s equa­
mation. These changes lead to uncertainties regarding the integrity of
tion is needed which describes the interactions between the fluid and
the reservoir rock and sealing cap rock. Moreover, geochemical re­
rock and can be written as depicted in Eq. (15):
actions may result in the dissolution of rock minerals and consequently
decreasing the reservoir rock and cap rock integrity leading to possible σ′ = σ – βp 15
gas leakage. Another consideration in the geomechanical aspects of UHS
is the pre-existing faults and fractures that might be reactivated due to where σ′ is the effective stress, σ is the total stress, P is the pore pressure,
the stress changes exerted by the cyclical injection and production of and β is the Biot coefficient (Gram et al., 2021). Hook’s Law is another
hydrogen (Kanaani and Sedaee, 2022; Aftab et al., 2022). This, conse­ fundamental equation utilised to model the geomechanical effects of
quently, leads to a higher risk of gas leakage and losses. Therefore, it is UHS. It describes the elastic deformation of the rock matrix by relating
crucial to evaluate the geomechanical effects of UHS on the reservoir the stress and strain and it can be written in the following form (Eq. (16))
formation.
σ = Eε 16
To model geomechanical changes during UHS it is important the
determine the stresses exerted on the reservoir rock as a result of the
where σ is the stress, ε is the strain, and E is the elastic modulus of the
hydrogen gas injection and production. Moreover, it is necessary to
material. In order to predict the onset of rock failure, Mohr-Coulomb
determine the rock’s behaviour as a result of these exerted stresses in
failure criterion can be used. This criterion relates shear strength of
terms of strength, strain and elasticity. One of the fundamental equa­
the rock to the normal stress and friction angle, and is given by:
tions used to model these geomechanical changes is the fluid conser­
vation of momentum equation which is given by Eq. 14 τ = c + σn tan(φ) 17

8
M. Saeed and P. Jadhawar
Fig. 6. Most important parameters and phenomena in modelling the geochemical aspects of UHS.
where τ is the shear stress, c is the cohesive strength of the rock, σn is the
normal stress, and φ is the angle of internal friction (Al-Awad, 2002). In
UHS modelling, the conservation of momentum equation is used to
calculate the pore pressure changes due to hydrogen injection and
withdrawal. These changes are then used as input to Biot’s equation to
calculate the resulting volumetric strain in the reservoir rock. The strain
is then used to calculate the stress changes using Hook’s law. The stress
changes are then compared to the strength of the rock, as described by
the Mohr-Coulomb criterion, to evaluate the potential for reservoir
failure.
Modelling the geomechanical effects of hydrogen gas injection,
production, and flow in porous media can be performed in different
manners. Two of the techniques to model these effects are one-way
coupling or two-way coupling. In one-way coupling, the fluid flow and
geomechanical simulations are done sequentially. Typically, the fluid
flow simulation is conducted first, generating changes in pore pressure.
Then, these changes are fed into the geomechanical model to evaluate
stress and deformation. One the other hand, two-way coupling refers to
the approach wherein changes in fluid pressures and saturations within
the porous medium directly influence the stress-strain behaviour of the
solid matrix, and reciprocally, the deformation or mechanical response
of the solid matrix concurrently affects the fluid flow paths, saturations,
and pressure distributions. Two-coupling ensures that the intricate in­
teractions between fluid dynamics and geomechanical responses are
captured in real-time throughout the simulation providing more accu­
rate results (Hawez et al., 2021). Modelling further processes or effects
that are co-dependent such as thermal changes and geochemical re­
actions can also be performed through fully coupled simulations.
Including such interconnected processes in fully coupled simulations
would provide more accurate results in modelling UHS. However, these
types of simulation techniques are much more computationally
compared to the one-way coupled technique due to the required itera­
tive technique.
9
Gas Science and Engineering 121 (2024) 205196
2.2.2. Analysis of previous studies’ findings
Bai et al. (Bai and Tahmasebi, 2022) developed a 3D coupled
hydro-mechanical model to evaluate the feasibility of UHS in a saline
aquifer at the Powder River Basin of Wyoming State. In their work, they
used the one-way coupling approach to simulate the hydromechanical
response of UHS in the aquifer, which assumed rock strain and stress are
affected by pore pressure without affecting the pore pressure back.
Hydrogen properties were calculated based on the Peng-Robinson
equation of state. They reported that a maximum of 75% hydrogen
could be recovered by the end of the third cycle. Moreover, the integrity
of the formation and caprock was evaluated using the Mohr-Coulomb
criterion which showed that UHS in the studied structure is geo­
mechanically safe. It was also shown that the possibility of failure de­
creases during the withdrawal period due to a drop in pore pressure and
an increase in the effective stress. Fang et al. (2022) performed nu­
merical simulations using the Thermo-Hydro-Mechanical (THM)
coupled simulator FLAC3D-TOUGH2MP to evaluate the applicability of
UHS in thin-bedded Anning salt mine (China). TOUGH2MP performs the
thermal and hydraulic simulation and FLAC3D performs the mechanical
calculations. They demonstrated that the coupled simulator does pro­
vide means to evaluate the complex pressure and temperature changes
for short term storage in salt cavern. Their work showed that the studied
potential salt cavern in Anning mine satisfies the criteria for safe
hydrogen storage which are no-spalling, no-creep failure, convergence
volume of salt cavern does not affect its storage capacity, and resultant
surface subsidence rate should not disturb the surface structures.
Bottcher et al. (Böttcher et al., 2017) carried out numerical model­
ling through openly available simulator OpenGeoSys to investigate the
thermo-mechanical behaviour during UHS in salt caverns. Their results
showed that large temperature amplitudes in the working gas may in­
crease the tensile stresses at the cavern boundary. They concluded that
reducing the number and frequency of storage cycles can decrease
temperature variations to avoid tensile failures. Kumar et al. (2021)
M. Saeed and P. Jadhawar
investigated the heterogeneity on the deformation of salt caverns and
quantify the state of stress around the caverns by developing an open
source 2D finite element simulator. They benchmarked their simulation
scheme with experimental data from published literature. Their results
showed that salt cavern creep (permanent deformation of solid rocks
under the influence of persistent stress) was a slow process and was not
affected by the short-term storage of hydrogen. Moreover, the hetero­
geneity of the salt cavern affects its elasticity differently based on the
type (potash, halite) and distribution of impurity. Where the presence of
heterogeneous layers was observed to result in much higher local de­
formations around the cavern compared to a homogeneous formation.
Coarita-Tintaya et al. (2023) evaluated the viability and risks associated
with utilising salt caverns for UHS. Through a numerical approach, their
research integrated hydromechanical characteristics of rock salt
encapsulating short- and long-term mechanical behaviours. Their model
encompassed elastoplastic, instantaneous damage mechanisms, as well
as primary, secondary, and tertiary creep phases. By applying this
model, the hydromechanical performance of vertical salt caverns was
evaluated for both shallow and deep hydrogen storage scenarios sub­
jected to cyclic injection and withdrawal. They reported that deeper
caverns exhibit heightened stability concerns, particularly under daily
operational cycles. Nevertheless, gas leakage remains limited due to
predominant diffusion transport, even with varying permeabilities.
Consequently, they concluded that the stability of salt caverns for green
hydrogen storage appears resilient under aggressive operational condi­
tions, boding well for the potential of hydrogen storage initiatives
(Coarita-Tintaya et al., 2023). A summary of the modelling studies on
geomechanics during UHS is presented in Table 3.
2.3. Hydrodynamic
This section covers hydrodynamic aspects of UHS, including funda­
mentals, modelling approaches, and findings from previous studies. The
main hydrodynamic phenomena observed during UHS can be sum­
marised as H2 lateral spreading, H2 viscous fingering, gas mixing, and
relative permeability and capillary pressure hysteresis trapping. These
phenomena are dictated by various operational conditions, and fluid
and rock properties such as fluids densities and viscosities, porosity,
permeability, relative permeability and capillary pressure, and forma­
tion geology. These properties and phenomena will be discussed in
detail over the next sections. Fig. 7 provides a summary of the primary
parameters and phenomena for hydrodynamic modelling.
Table 3
A summary of studies on geomechanical modelling of UHS.
Reference Scope Storage formation
type
Bai and Tahmasebi Investigate the hydro-mechanical Saline aquifer at the
(2022) aspects of UHS Powder River Basin of
Wyoming State
Böttcher et al. Assess the geomechanical aspects of Salt cavern
(2017) UHS in salt caverns
Fang et al. (2022) Assessed geomechanical effects during Salt cavern
UHS
Kumar et al. (2021) Investigated the heterogeneity on the Salt cavern
deformation of salt caverns and
quantify the state of stress around the
caverns
Coarita-Tintaya Evaluated the integrity salt caverns Salt cavern
et al. (2023) during UHS
10
Gas Science and Engineering 121 (2024) 205196
2.3.1. Fundamentals and modelling approaches
Hydrodynamics plays a critical role in understanding fluid flow
phenomena within storage formations and is a fundamental aspect of
UHS (see Fig. 8). The hydrodynamic behaviour of a fluid in a storage
formation is governed by a complex interplay of fluid properties, rock
properties, and rock-fluid properties under the prevailing reservoir
conditions. In particular, fluid viscosity, density, interfacial tension
(IFT), solubility, and diffusivity are important fluid properties that
impact fluid flow in storage formations (Muhammed et al., 2022). Rock
permeability and porosity are crucial rock properties that determine the
ease of fluid flow within the rock and the rock’s storage capacity,
respectively. Furthermore, the interactions between rock and fluids
under the prevailing reservoir conditions are also critical in determining
the efficiency of storage and withdrawal processes (Zivar et al., 2021a;
Muhammed et al., 2022; Hashemi et al., 2021a). The rock-fluid prop­
erties of wettability, capillary pressure, and relative permeability also
significantly influence the amount of hydrogen that can be stored and
recovered and the ease of injection and production (Hashemi et al.,
2021b; Ali et al., 2021). As a result, the integration of the aforemen­
tioned properties into hydrodynamic models is essential for accurately
predicting UHS performance, making hydrodynamics arguably the most
important aspect of UHS modelling.
2.3.1.1. Fluid properties. The most important fluid properties consid­
ered in UHS hydrodynamics modelling are viscosity and density. Vis­
cosity is the measure of a fluid’s resistance to flow. Hydrogen’s viscosity
is very low compared to other fluids and is 0.0189 cP at standard con­
ditions (Muhammed et al., 2022). This results in challenges during the
injection of hydrogen gas into the geological formation. Due to its high
mobility, hydrogen gas tends to escape and spread laterally in the
reservoir leading to undesirable stored gas losses (Heinemann et al.,
2021a). Viscosity can be calculated using the Peng-Robinson EOS which
takes into account the intermolecular interactions between the gas
molecules (Nasrifar, 2010). Hydrogen gas has a very low density, and it
affects the fluids distribution of fluids inside the geological structure
leading the hydrogen gas to rise to the top of the structure. Hydrogen gas
density and density contrast between the underground fluids become
important in the absence of the viscous forces (resulting from injection
and production operations) (Nasrifar, 2010). H2 density can also be
calculated by first obtaining the molar volume (Vm ) from the
Peng-Robinson EOS and then utilising the ideal gas law:
Vm = V/n = RT/P 18
Approaches and Software Remarks
packages
3D coupled hydro- Maximum of 75% hydrogen could be recovered by the
mechanical model end of the third cycle the studied structure is
Peng-Robinson equation of geomechanically safe
state The possibility of failure decreases during the withdrawal
period
OpenGeoSys large temperature amplitudes in the working gas may
increase the tensile stresses at the cavern boundary
Thermo-Hydro-Mechanical Convergence volume of salt cavern does not affect its
(THM) coupled simulator storage capacity, and resultant surface subsidence rate
FLAC3D-TOUGH2MP should not disturb the surface structures.
Developed an open source Salt cavern creep was a slow process and was not affected
2D finite element simulator by the short-term storage of hydrogen.
Heterogeneity of the salt cavern affects its elasticity
differently based on the type (potash, halite) and
distribution of impurity.
Numerical modelling Deeper caverns exhibit heightened stability concerns
compared to shallower, particularly under daily
operational cycles. Nevertheless, gas leakage remains
limited due to predominant diffusion transport, even
with varying permeabilities.
M. Saeed and P. Jadhawar Gas Science and Engineering 121 (2024) 205196

Fig. 7. Most important parameters and phenomena in modelling the geochemical aspects of UHS.

Fig. 8. Reservoir mechanisms during UHS (from Heinemann et al. (Heinemann et al., 2021a)).

where P is the pressure, V is the volume, n is the number of moles of gas, 2.3.1.2. Rock properties. Permeability and porosity are considered the
R is the gas constant, and T is the temperature. most important rock properties dictating the hydrodynamics of fluids
flow in porous media (see Fig. 9). Permeability is the measure of the
rock’s ability to permit fluid flow through it. Permeability is calculated

11
M. Saeed and P. Jadhawar
Fig. 9. Representation of rock porosity and permeability (Joseferd, 2015).
using Darcy’s law and it depends on the properties of the porous media,
such as pore size, shape, and connectivity. Darcy’s Law is the basic
equation controlling fluid flow through porous media. It connects the
pressure gradient and permeability of the porous media to the fluid flow
rate (Dake, 1978). Darcy’s law, which describes hydrogen flow in porous
media, may be expressed as
Q = − kA (P / μL) 19
where P is the pressure gradient resulting in the flow, Q is the fluid’s
flow rate, k is the permeability of the porous medium, A is the cross-
sectional area of the media, μ is the viscosity of fluid, and L is the
length of the porous media. Permeabilities depend on the rock type and
the depositional environment through which the rock was formed. High
permeability and homogenous reservoirs are desirable as they are ex­
pected to improve the hydrogen gas’s injectivity and producibility.
Another important rock property dictating the hydrodynamics during
UHS is the rock’s porosity which is the measure of the rock’s ability to
store fluids within its pores. Absolute porosity is a percentage of the pore
space available in a rock to the bulk volume of the rock (Dake, 1978).
Effective porosity is the percentage of connected pores’ space to the
rock’s bulk volume, and it is the porosity that is considered in modelling
hydrogen gas, and generally fluids, flow in porous media.
2.3.1.3. Rock-fluid properties. Relative permeability and capillary
pressure are two essential properties in UHS that influence the flow and
distribution of fluids in porous media. Relative permeability is a mea­
sure of how easily a fluid can flow through a porous medium compared
to another fluid. It is defined as the ratio of the effective permeability of
a fluid to its absolute permeability (Pan et al., 2023). It can be calculated
using several models such as Stone I, Stone II, and Modified
Brooks-Corey Models. The Modified Corey Model can be used to esti­
mate relative permeability in hydrogen-water systems, where the porous
media is saturated with the two immiscible fluids (Ahmed, 2010). The
model assumes that the relative permeability of each fluid phase is a
function of its own saturation and the saturations of the other two phases
(wetting and non-wetting phase), and can be written in the general form
of equations Eq. 20 to Eq. (21):
( )n1
Sw − Sw,ir
0
kr,w = kr,w 20
1 − Sw,ir − Snw,r
( )n 2
1 − Sw − Snw,r
0
kr,nw = kr,nw 21
1 − Sw,ir − Snw,r
where Sw and Sw,ir the water saturation and irreducible water saturation,
Snw,r is the residual saturation of the non-wetting phase (hydrogen gas in
this case), n1 and n2 are empirical exponents that depend on the prop­
12
Gas Science and Engineering 121 (2024) 205196
erties of the reservoir and the fluids, and k0r,w and k0r,nw are the end point
relative permeabilities for the wetting and non-wetting phase, respec­
tively. Capillary pressure is the pressure difference between two
immiscible fluids, such as water and hydrogen, in a porous media. It is
an important parameter for predicting fluids distribution in a porous
media, and it can be calculated using the Young-Laplace equation (Eq.
22).
Pc = 2σIFT cos θ / r 22
where Pc is the capillary pressure, σ IFT is the interfacial tension between
the two fluids, θ is the contact angle between the fluid and the solid
surface of the porous media, and r is the radius of curvature of the fluid-
fluid interface (Dake, 1978; Ahmed, 2010). Relative permeability and
capillary pressure determine the producible hydrogen gas saturation
during withdrawal stage.
A few measurements of H2-water relative permeability data have
been recorded in the literature. Fig. 10 shows the drainage and imbi­
bition relative permeability curves for H2-water system reported by
Boon and Hajibeygi (2022) and Lysyy et al. (2022a). It is noted from that
a pronounced difference between the drainage and imbibition curves
end points exists for both datasets which is a result of the relative
permeability hysteresis phenomena. The differences in the residual gas
saturation range approximately between 0.28 and 0.42 which then
translates to residual gas trapped between the injection and withdrawal
phases. The residual trapping of hydrogen gas during the withdrawal
phase can be seen on a micro level in Fig. 11. Other forms of hydrogen
trapping during UHS are structural trapping due to the entrapment of
hydrogen gas within the structures of rocks and dissolution trapping as a
result of hydron gas dissolution in water. Hydrogen entrapment as a
result of geochemical reactions, biotic and abiotic, is another potential
mechanism for losing hydrogen gas during UHS as discussed previously.
The hysteresis between the drainage and imbibition curves of rela­
tive permeability result in the entrapment of residual hydrogen gas as a
result of the discontinuity in the hydrogen gas phase while the con­
nected hydrogen phase is being produced. This is particularly important
in underground hydrogen storage because of the cyclic nature of the
operation such that successive injection and withdrawal of hydrogen to
and from the porous media result in change in the flow behaviour and
consequently relative permeability curves. The hysteresis effects can be
assessed by evaluating the amount of the non-wetting phase, here being
hydrogen, that stays trapped during a shift in flow from drainage to
imbibition. The Land model (Land, 1968) is frequently utilised to
investigate this trapping behaviour. This model creates an empirical link
between the residual hydrogen saturation SH2 ,r (hydrogen quantity
trapped when capillary pressure is null) and the hydrogen saturation at
the instant of flow inversion, SH2 ,i . The Land model for H2-water system
M. Saeed and P. Jadhawar Gas Science and Engineering 121 (2024) 205196

Fig. 10. H2-water relative permeability for drainage and imbibition (from Boon and Hajibeygi (Boon and Hajibeygi, 2022) and Lysyy et al. (Lysyy et al., 2022a)).

Fig. 11. Residual entrapment of hydrogen gas during UHS in an aquifer (modified after Lysyy et al. (Lysyy et al., 2022b)).

can be written in the following set of equations Eq. 23 to Eq. 25 total trapping, whereas C nearing 1 indicates negligible trapping. For
any intermediate value of C, the residually trapped hydrogen gas pro­
SH* 2 ,i
SH* 2 ,r = 23 portion escalates with an increase in the initial gas saturation. To sup­
1 + CSH* 2 ,i plement the Land model, Killough (1976) introduced an interpolation
method for identifying scanning curves, which serve as intermediate
1
QC = 24 curves for forecasting the relative permeability curve and its endpoints
SH2 ,r + CSH* 2 ,i (Eq. (26) to Eq. (27))

Sw SH2 − SH2 ,i
Sw* = 25 SH2 ,r = Smax,H2 ,r + ( ) 26
1 − Sw,irr 1 + C SH2 − SH2 ,i

where, S*H2 = 1 − S*w and C is a positive constant, known as the Land C=

1
×
1
27
coefficient, which is determined experimentally. When C = 0, it signifies Smax,H2 ,r − SH2 ,i Smax,H2 − SH2 ,i

13
M. Saeed and P. Jadhawar
By applying a parametric interpolation method, the complete imbi­
bition relative permeability curve for the hydrogen-water system can be
derived. This method entails interpolating between the point of flow
inversion SH2 ,i and the corresponding residual saturation SH2 ,r .
2.3.1.4. Fluid-fluid properties. IFT is a measure of the attractive forces
between the molecules at the interface of two immiscible fluids, and it
can affect the capillary pressure and the relative permeability of the
fluids within the porous media. H2-water IFT can be directly measured
in the laboratory (Hosseini et al., 2022), calculated as a function of
capillary pressure using the Young-Laplace Equation (Eq. 22) or
modelled through molecular dynamic simulations (van Rooijen et al.,
2022). Another fluid-fluid phenomenon that affects the hydrodynamics
of UHS is the hydrogen gas’s molecular diffusivity. Diffusivity describes
the ability of hydrogen gas molecules to dissipate through a porous
media (Charlet et al., 2017). Molecular diffusion is usually calculated
using Fick’s law which can be written as follows:
J = ρw Dkw ∇ckw 28
where Dkw is the effective diffusion coefficient of component k, c is the
mole fraction of component k (Hagemann et al., 2015).
Hydrogen dissolves in water in small amounts in it depends on
exerted pressure, temperature, and water salinity. Many experimental
studies measured the solubility of hydrogen in water in different con­
ditions (Wiebe and Gaddy, 1934; Chabab et al., 2020). Hydrogen gas
solubility in water can also be calculated using Henry’s Law as a function
of the pressure’s partial pressure, Pp , and can be expressed as follows:
C = kH Pp 29
where C is the concentration of the gas in the liquid, and kH is the
Henry’s law constant, which is specific to the gas and the liquid. The
value of kH for hydrogen gas in water can be obtained from experimental
measurements or from thermodynamic databases such as the NIST
Chemistry WebBook and the Dortmund Data Bank (Sass, 2013).
2.3.2. Fundamental equations
The fundamental equations required to model hydrogen flow in
porous media are mass and energy conservations equations, flow
equation (Darcy’s Law), and an equation of state. Two of the most
widely used Equation of States for PVT behaviour modelling in standard
reservoir simulators are Peng Robinson EoS (Eq (12)) and SoaveRedlich-
Kwong EoS. Another important Equation of State that can be imple­
mented for UHS is GERG-2008 equation of state (EOS) named after the
group that developed it European Gas Research Group. However, most
of the standard reservoir simulators do not include the option to utilise
GERG-2008 EoS. The decision to adopt an EoS for modelling UHS de­
pends on the EoS that best predicts the hydrogen gas properties at the
desired application conditions. The conservation of mass equation states
that the rate of change of mass within a control volume is equal to the
net mass flow rate into or out of the control volume and can be written
as:
∂(φSα ρα ) / ∂t + ∇ • (qα ) = Qα 30
where, φ is the porosity, Sα is the saturation of the α phase (α = water,
oil, gas), ρα is the density, qα is the fluid flowrate, Qα is the source/sink
term, and α is the phase (water, oil, or gas). Similar to the conservation
of mass, the conservation of energy equation states that the rate of
change of energy within a control volume is equal to the net energy flow
rate into or out of the control volume and can be written as:
ρα cpα (∂T / ∂t + div(qα T)) = ∇.(kα ∇T) + Hα 31
where ρ is the density, cp is the specific heat capacity, T is the temper­
ature, k is the thermal conductivity of the porous medium, H is the rate
of heat generation or consumption within the control volume. Darcy’s
14
Gas Science and Engineering 121 (2024) 205196
law (Eq. (32)) describes the fluid flow in porous and can be expressed in
a more generic way as (Hagemann et al., 2015):
vα = λα .(∇Pα − ρα g) 32
where v is the velocity, and λ is the mobility of phase α. In addition to the
above fundamental equations, the sum of saturations and sum of con­
centrations in a phase both are equal to one. Utilising the above-
mentioned equations, the compositional transport equation for a
component γ in a multiphase 3D system and can be written as:
( )/ ( ) ( )
∂ φSα ρα cγ ∂t + ∇ φSα ρα cγ vα = ∇ • Dαeff ∇cγ + qα 33
where cγ is the concentration of component γ. Eq. (33) is a partial dif­
ferential equation (PDE) that can be solved numerically. In order to
solve the transport equations, the first step is to discretise them in space
and time. After discretising the transport equations in space and time,
the initial and boundary conditions are set, and the equations are solved
using numerical techniques such as the Gauss-Seidel method and the
Jacobi method (Chen, 2007).
Dimensionless numbers are often used in reservoir simulations to
characterize the behaviour of reservoirs and to facilitate the analysis of
fluid flow through porous media. A group of key dimensionless numbers
for reservoir modelling is defined and presented in Table 4.
Multiple software packages have been employed for simulating the
hydrodynamic flow of H2 during UHS. Among these packages, DuMux,
SLB’s Eclipse, CMG-GEM, COMSOL, and TOUGH are the most frequently
used (Computer Modelling Group, 2022; Schlumberger Ltd, 2014;
Pruess et al., 1999; Dickinson et al., 2014; Flemisch et al., 2011). Most of
these software packages employ similar principles of fluid flow through
porous media. DuMux is an open-source simulator that models multi­
phase multicomponent transport in porous media. It is integrated into
Distributed and Unified Numerics Environment (DUNE) and solves the
pressure, saturation, and flow equations using either the coupled fully
implicit approach or the decoupled semi-implicit approach. Being
open-source, DuMux provides an opportunity to adjust the flow equa­
tions and couple them with various equations of state to appropriately
simulate different flow applications (Flemisch et al., 2011).
SLB’s Eclipse 100 and 300 are reservoir simulation tools that are
typically employed for dynamic modelling of oil and gas reservoirs.
Eclipse 100 models non-compositional fluid flow in porous media using
a black oil model, which simulates the flow of phases as a bulk without
considering their respective components. These phases are oil phase,
water phase, and gas phase (Schlumberger Ltd, 2014). SLB’s composi­
tional simulator (Eclipse 300) and similarly CMG-GEM account for
phase behaviour changes during processes where changes in fluid
compositions are expected to take place. In addition to Darcy’s equa­
tions, compositional simulators use mass balance for each component
and the appropriate equation of states to model multiphase fluid flow in
porous media. The most widely used equation of state in petroleum
simulation software is the Peng-Robinson EOS (Computer Modelling
Group, 2022; Schlumberger Ltd, 2014).
TOUGH (Transport of Unsaturated Groundwater and Heat) is
another simulation suite that numerically models fluid flow in porous
media and can be used for gas injection and reactive transport in porous
media, which is also suitable for modelling the injection of H2 and
cushion gases in porous media (Pruess et al., 1999). COMSOL Multi­
physics has also been utilised for modelling UHS. It is a simulation
environment where multiple scientific models are present to simulate a
variety of processes, such as acoustics, fluid flow, and chemical reactions
(Dickinson et al., 2014). A comparison of the popular simulators used for
modelling fluid flow in porous media and UHS is presented in Table 5.
2.3.3. Analysis of previous studies’ findings
Hagemann et al., 2015, 2016 developed a mathematical model and
implemented it numerically in DuMux to investigate the hydrodynamic
effects of UHS. Their findings indicated that gravitational forces were
M. Saeed and P. Jadhawar
Table 4
A group of key dimensionless numbers in fluid flow in porous media and
reservoir engineering.
Dimensionless Equation Components Significance and
Number Application
Capillary μH2 is the viscosity of The ratio of viscous
Ca =
Number (Ca) hydrogen, V is the forces to capillary forces
μH2 V
σ
Darcy velocity, and σ is in a fluid flow through
the IFT between porous media. It gauges
hydrogen and water the relative importance
of viscous forces to
capillary forces. A high
Capillary Number
indicates dominance of
viscous forces, which is
essential in enhanced
recovery processes. In
reservoir simulation,
this number can help in
understanding the
displacement efficiency
of hydrogen during
injection and
withdrawal.
Darcy Number k k is the permeability of It is used in reservoir
Da =
(Da) L2 the medium, and L is a models to characterize
characteristic length. the porous media and
predict how easily
fluids can traverse
through it, thereby
affecting storage and
extraction rates of
hydrogen
Mobility Ratio λH2 The ratio of the mobility
M =
(M) λw of the displacing fluid to
the mobility of the
displaced fluid and for
the injection phase
during UHS. It
compares the mobility
of hydrogen to that of
water within the
reservoir, which is
crucial for the
displacement processes.
Fractional
f H2 =
where λH2 and λw are The fraction of total
Flow (f) the mobilities of the flow due to one phase. It
λH 2 hydrogen and water can also be written for
λw + λH 2 phases, respectively. water by replacing the
hydrogen mobility in
the numerator with
water mobility.
Leverett J- where Pc is the This function is a
function (J) J = capillary pressure, and dimensionless form of
√̅̅̅ r0 is the average pore the capillary pressure
Pc r0
k radius curve. It used in
σ φ reservoir simulation to
characterize capillary
pressure curves which
affect the distribution
and movement of
hydrogen and water in
the reservoir.
Peclet Number UL where U is the velocity Represents the ratio of
Pe =
(Pe) D of the fluid, L is a advective to diffusive
characteristic length, transport. It is utilised
and D is the diffusivity to analyse heat and
mass transfer effects,
which can impact the
behaviour of hydrogen
movement and mixing
within the reservoir.
Courant where U is the velocity This number is used in
Co =
Number (Co) of hydrogen or water, setting time step sizes in
UΔt Δt is the time step size, numerical simulations
Δx and Δx is the spatial to ensure stable and
step size accurate solutions,
especially in transient
simulation scenarios.
15
Gas Science and Engineering 121 (2024) 205196
dominant at low injection rates, resulting in uniform water displace­
ment, while at higher injection rates, viscous forces became more
dominant, leading to unstable water displacement and lateral gas
fingering. Stratified aquifers were found to limit the risk of gas loss due
to lateral spreading. However, this operation was affected by the ve­
locity at which gas rises toward the cap rock seal and requires gas in­
jection into lower structures. Furthermore, the authors coupled DuMux
with biochemical modelling to investigate the biochemical effects on
UHS in a depleted gas reservoir. They observed that microbial growth
and decay rates significantly influenced the process of hydrogen con­
version into methane and ignoring methanogenesis led to significant
differences in the results. Feldmann et al. (2016) also used numerical
modelling through DuMux to investigate hydrodynamics and gas mixing
during UHS in a depleted gas reservoir. They concluded that gravity
override and viscous fingering phenomena hindered the UHS process in
aquifers but had a limited effect on UHS in gas-saturated reservoirs.
Additionally, mechanical dispersion amplified the gas mixing process
more than molecular diffusion.
Pfeiffer et al. (Sainz-Garcia et al., 2017a; Pfeiffer and Bauer, 2015;
Pfeiffer et al., 2016, 2017) utilised SLB’s compositional simulator
Eclipse 300 to study the possibility of hydrogen storage in the North
German Basin. They found that storage performance increased with the
number of cycles, and shut-in periods during the initial filling period
improved injectivity by lowering reservoir pressures during these pe­
riods. In another scenario, they used five wells for the injection and
withdrawal processes, with nitrogen gas injected first to act as the
cushion gas. The results confirmed that the storage performance
increased with the increased number of storage cycles, and hydrogen
loss due to dissolution in connate water was 1.75% of the initial
hydrogen storage over a four-year period. In 2017, Pfeiffer et al.
extended their efforts by utilising numerical simulation and facies
modelling using ECLIPSE 300 simulator to investigate the possibility of
UHS in the North German Basin. They showed that on average, UHS can
sustain a continuous power output of 245 MW for one week using five
storage wells. Enigbokan et al. (2021) used numerical simulation
through Eclipse 300 to assess the feasibility of UHS in a depleted gas
field, the Viking field in the UK. Their results showed that large-scale
seasonal storage can be achieved in the studied field. However, their
model did not include hydrogen loss due to solubility, geochemical re­
actions, or diffusion effects, and can be considered as a best-case sce­
nario in the absence of these effects. Lysyy et al. (2021) used the black
oil simulator Eclipse 100 to model hydrogen gas storage in the Norne
hydrocarbon field (Norway). Their modelling results showed that UHS
in the gas zone was the favoured option, resulting in a hydrogen re­
covery of 87% compared to 77% and 49% hydrogen recovery from oil
and water zones, respectively. Injection of formation gas as a cushion gas
resulted in improved hydrogen recovery but also affected the purity of
the produced hydrogen gas due to gas mixing.
Luboń and Tarkowski (2020) utilised PetraSim-TOUGH2 coupled
with the EWASG module to simulate seasonal hydrogen storage in a
deep aquifer structure in Poland. They assumed one well positioned at
the summit of the structure was used for both injection and production
and that the fracturing pressure and capillary entry pressure would not
be exceeded. Their modelling results indicated that water coning and
excessive production are the main obstacles for UHS in the aquifer. They
also found that hydrogen recovery depends on the initial injection
period, where longer injection periods result in lower hydrogen recov­
ery. For example, they found that a 36-month period will result in a
17.75% hydrogen recovery compared to 25.41% recovery in a 24-month
period. Luboń and Tarkowski (2021) compared the storage capacity of
carbon dioxide and hydrogen in deep aquifer anticline. Their results
showed that the structure can store up to 1 million tons of carbon di­
oxide compared to 4000 tons of hydrogen over a 31-year period. This
was due to the higher density of carbon dioxide, which allows a higher
mass of carbon dioxide to be stored under the same pressure. They also
found that altering the threshold capillary pressure of the cap rock has a
M. Saeed and P. Jadhawar Gas Science and Engineering 121 (2024) 205196

Table 5
Basic comparison between the popular simulators used for modelling fluid flow in porous media and UHS.
Criteria E100, E300 Intersect CMG (IMEX, TOUGH MRST DuMux OPM OpenGeoSys COMSOL
GEM, STARS)

Main Application Reservoir Reservoir Reservoir Reactive Reservoir Subsurface Reservoir Environmental Multiphysics
Simulation Simulation Simulation Transport Simulation Flow Simulation Simulations
Spatial Finite Finite Finite Finite Volumes Finite Finite Finite Finite Volumes Finite
Discretization Differences Differences Differences Volumes Volumes Volumes Elements
Temporal Implicit Implicit Implicit Implicit/ Implicit/ Back Eueler Back Eueler Various Implicit
Discretization Explicit Explicit
Compositional E300 Yes Yes (GEM) Yes Yes Yes Yes Yes Yes
Simulation
Geochemical Limited Limited Limited Yes Limited Limited Limited Yes Yes
Modelling
Geomechanical Limited Yes Limited Limited Limited Yes No Yes Yes
Modelling
Coupled No Yes Limited No Limited Yes No Yes Yes
Geomechanical
Modelling
Thermal Modelling E300 Yes Yes (STARS) Yes Limited Limited Limited Yes Yes
Computational Moderate to High Moderate to Moderate to Moderate Moderate Moderate Moderate to High High
Requirements High High High
Parallel Computing Yes Yes Yes Limited Limited Yes Yes Yes Yes
Licensing Commercial Commercial Commercial Commercial/ Open- Open- Open- Open-Source Commercial
Open-Source Source Source Source
Programming C++ C++ C++ Fortran/C++ MATLAB C++ C++ C++ Java
Language

significant impact on the modelling results and suggested that more
studies are required to fully understand this effect. Mahdi et al. (2021a)
used TOUGH2 coupled with the EWASG equation of state to evaluate
UHS in a heterogeneous sandstone reservoir. Their results showed that
higher injection rates lead to increased hydrogen leakage in the absence
of a sealing caprock and increases hydrogen losses due to undesirable
gas fingering. TOUGH was utilised by Huang et al. (2023) to understand
the UHS viability in a depleted gas reservoir. Their simulation encom­
passed a hydrogen injection-idle-withdrawal operation in an anticline
reservoir. They reported that hydrogen leakage through cap-rock and
aqueous phase losses (1%) was minimal. Yet, 73% hydrogen recovery
without cushion gas was observed. They investigated the utilization of
nitrogen and carbon dioxide cushion gases and reported an improved
hydrogen recovery to 91% and 81%, respectively, with nitrogen
exhibiting superior performance, enhancing both the hydrogen recovery
factor and gas purity.
Heinemann et al. (2021b) investigated the role of cushion gas during
UHS in saline aquifer anticline using the commercial compositional
simulator CMG-GEM at varying geological parameters (reservoir depth,
trap shape, and reservoir permeability). They concluded that cushion
gas does not expand the storage capacity, but it can be used to exploit the
existing capacity more efficiently. The cushion gas required depends on
the desired amount of working gas, so studies should be conducted early
in projects to determine cushion gas requirements and avoid unnec­
essary investment losses. Moreover, their simulation results showed that
hydrogen storage in geological structures at greater depths with higher
reservoir permeabilities requires lower amounts of cushion gas. They
also concluded that tighter anticlines make the injection process more
difficult without affecting the production, and more open anticlines
result in increasing water production. Kanaani et al. (2022) conducted a
study to investigate the impact of various types of cushion gases on the
UHS in depleted oil reservoirs using the CMG-GEM simulator. The au­
thors observed that UHS without the use of cushion gas resulted in up to
15.5% losses in the stored hydrogen. Methane was found to be the most
effective cushion gas, as it resulted in higher hydrogen recovery rates
compared to nitrogen and carbon dioxide. The authors attributed this to
the lower molecular weight of methane, which resulted in a smaller
density contrast relative to hydrogen. The simulation runs also showed
that increased rates of hydrogen injection in short intervals could limit
the effects of gravity segregation due to improved sweeping of reservoir
fluids. The impact of operational parameters and rates on geologic
hydrogen storage was also evaluated by Jadhawar and Saeed (2023a)
utilising CMG-GEM, where they studied the impact of injection rates,
production rates, shut-in periods, and frequency and length of storage
cycles. Their work showed that hydrogen recovery efficiency decreased
as the injection and production rates increased and that longer and less
frequent storage cycles were less beneficial in terms of hydrogen re­
covery. Their results also showed that the presence of a shut-in period
between the injection and production stages allows hydrogen gas to
escape away laterally from the operating well reducing the hydrogen
retrieval.
Wang et al., 2022a, 2022b utilised the CMG-GEM simulator to
investigate the applicability of the scaling theory to hydrogen and car­
bon dioxide flow inside a hypothetical reservoir. They found that
dimensionless groups such as aspect ratio, gravity/viscous ratio, and
gravity/capillary ratio could be used to correlate the hydrogen and
carbon dioxide flow between different scales. However, heterogeneous
reservoirs may not be optimal for UHS due to a drop in the purity of
produced hydrogen, especially at higher injection and production rates.
The simulation results also showed that sufficient quantities of cushion
gas could limit the effects of capillary pressure on the gas flow behav­
iour. The authors also investigated hydrogen and carbon dioxide flow
inside a heterogeneous reservoir and the effect of carbon dioxide solu­
bility in water on the process. Their results showed that due to CO2
solubility in brine, the volume of injected carbon dioxide needed to be
increased by approximately 30% to result in the same displacement
effect.
Delshad et al., 2022a, 2022b compared UHS in aquifers to depleted
hydrocarbon reservoirs using CMG-GEM. They found that both geolog­
ical storage sites possessed significant hydrogen storage and deliver­
ability capacity, with the cushion gas requirements being higher in the
aquifer case. Pre-existing wells and infrastructure in depleted hydro­
carbon fields presented a technical and economic advantage compared
to the aquifer case. The use of pump wells was also found to be beneficial
in improving the capacity of hydrogen storage in both aquifer and
depleted hydrocarbon reservoirs due to an improved voidage balance
inside the reservoir between the produced fluids through pump wells
and the injected hydrogen gas. Additionally, the inclusion of the effect of
hysteresis in relative permeability curves led to an increase in the
injectivity and working capacity of their studied storage aquifer. Their
numerical simulations showed that molecular diffusion and hydrogen
solubility in formation brine had a negligible impact on the final results.

16
M. Saeed and P. Jadhawar
Jadhawar and Saeed, 2023b, 2023c investigated the effects of reservoir
mechanisms and parameters on the performance of UHS in a deep North
Sea aquifer employing CMG-GEM for numerical simulations. Their re­
sults showed that with an increased number of storage cycles, there is an
upward trend in the efficiency of hydrogen recovery. Moreover, the
heterogeneity in aquifer permeability was found to be a pivotal deter­
minant, in defining the magnitude of hydrogen production. In homo­
geneous aquifers, the cumulative hydrogen production is quantified at
33.53 Bscf, contrasting with 30.38 Bscf in heterogeneous scenarios, thus
showcasing a 7% amplified hydrogen recovery in the former case. They
also reported that the incorporation of relative permeability hysteresis
introduces a tangible impact, leading to diminished volumes of both
injected and recoverable hydrogen—specifically, a decrease of 0.5 Bscf
in injected hydrogen volume and 1.24 Bscf in recoverable hydrogen
volume. They also noted that the cumulative influence of hydrogen
solubility and diffusion, though individually subtle, assumes greater
significance when interwoven with hysteresis within the analytical
framework. Saeed et al. (Saeed and Jadhawar, 2023; Saeed et al., 2023b)
utilised CMG-GEM to understand the impacts of cushion gas type and
physical properties on the efficiency of UHS. Their work showed that the
density of utilised cushion gas plays an important role in determining
the hydrogen recovery efficiency and that a lighter cushion gas was
more beneficial in terms of hydrogen recovery. Thus, CH4 gas was more
efficient as a cushion gas compared to N2 and CO2. They also concluded
that the utilization of a cushion gas was more favourable for hydrogen
recovery compared to operating without a cushion gas. Their results also
determined that the purity of recovered hydrogen gas was relatively
high at the start of each withdrawal phase in a storage cycle and reduced
as the withdrawal continued. However, the overall purity improved with
the cycles as the saturation of hydrogen gas near wellbore increased.
In another study, Cai et al. (2022) developed the simulation capa­
bilities of PSFLOW (General Purpose Subsurface Flow Simulator) for
modelling hydrogen and gas mixture storage in salt caverns, aquifers,
and depleted oil and gas reservoirs. The developed simulator was
benchmarked against other simulators and experimental results from
published literature. The authors observed that the developed simulator
was capable of modelling non-isothermal, multiphase, and multicom­
ponent flow in H2-water, H2–CH4-water, and H2–CH4–CO2-water sys­
tems at high pressure and temperature. However, their work did not
include geochemical and geomechanical effects.
Lubon and Tarkowski (Luboń and Tarkowski, 2023) conducted a
study examining how the duration of the initial filling phase affects the
storage of hydrogen in deep aquifers. Their results showed that
increasing the length of the initial filling period contributed to greater
operational capacity and better overall performance of the storage sys­
tem. A similar observation was made by Abdellatif et al. (2023a), who
confirmed that extending the initial filling period increased the amount
of hydrogen gas that could be recovered. They also discovered that
raising the rate at which hydrogen was injected actually reduced the
efficiency of hydrogen recovery. Specifically, when injection rates were
increased from 20 million standard cubic feet per day (MMscf/d) to 30
MMscf/d, the recovery rate dropped from 70% to 54%. This decrease in
hydrogen recovery due to higher injection rates was also supported by
Mahdi et al. (2021a), who identified factors like gas fingering and the
displacement of existing reservoir fluids as causes for this outcome.
Several studies have evaluated the effects of injection and production
rates on hydrogen recovery efficiency during underground hydrogen
recovery, with a consensus that increasing hydrogen rates negatively
affects hydrogen recovery efficiency (Hagemann et al., 2015, 2016;
Mahdi et al., 2021a; Wang et al., 2022a, 2022b). This negative impact is
attributed to the dominance of viscous forces at higher operating rates
compared to lower operating rates, where gravity forces are dominant.
Viscous fingering and lateral spreading, both of which result from the
low viscosity of hydrogen compared to other fluids in storage, are pro­
moted by the dominance of viscous forces. However, Kanaani et al.
(2022) have shown that higher injection rates combined with shorter
17
Gas Science and Engineering 121 (2024) 205196
injection intervals lead to improved hydrogen recovery efficiency due to
the minimized effects of gravity segregation resulting from improved
sweeping of reservoir fluids.
The expected recovery efficiency of hydrogen during UHS reported
in the various modelling studies varies significantly from approximately
30%–90%. The hydrodynamic behaviour of hydrogen and other fluids
during storage is the primary factor determining hydrogen recovery
efficiency, and this behaviour depends on various operational parame­
ters, reservoir-related properties, and mechanisms, as explained and
illustrated in Fig. 8.
Cushion gas plays a beneficial role in UHS and improving the
hydrogen recovery efficiency compared to the absence of cushion gas
(Sainz-Garcia et al., 2017a; Heinemann et al., 2021b; Kanaani et al.,
2022; Abdellatif et al., 2023a). The performance of hydrogen storage
depends on the type of cushion gas used, and CH4 and N2 have been
found to be the most efficient due to their lower density contrast with
hydrogen gas (Kanaani et al., 2022; Abdellatif et al., 2023a). This leads
to decreased viscous fingering and lateral spreading, which would occur
between hydrogen gas and resident reservoir fluids such as water in the
absence of cushion gas. The role of cushion gas is crucial in UHS and
evaluating cushion gas volume requirements is essential before under­
taking any UHS project.
Among the various reservoir-related mechanisms and phenomena
investigated in the literature, relative permeability hysteresis is one of
the most important factors affecting the performance of an UHS process
(Jadhawar and Saeed, 2023b; Wang et al., 2023a). The hysteresis of
relative permeability between hydrogen gas and water in porous media
indicates different behaviours of the two fluids during imbibition or
drainage processes. This phenomenon results in the retention of some
hydrogen gas after injection due to its inability to flow back with the rest
of the gas phase during withdrawal. The altered flow behaviour dis­
connects some of the hydrogen gas phase within the rock, trapping it
inside the reservoir. Both Delshad et al. (2022a) and Bo et al. (2023)
found that including H2-water relative permeability hysteresis in the
simulation improves hydrogen gas injectivity but reduces hydrogen gas
recovery efficiency. The increase in injection is attributed to the
improved water permeability compared to hydrogen, resulting in easier
displacement of water by hydrogen. However, hysteresis leads to higher
quantities of hydrogen gas becoming trapped inside the reservoir,
leading to lower hydrogen production. Hydrogen gas solubility and
diffusion were found to be negligible within the context of UHS (Pfeiffer
et al., 2017; Delshad et al., 2022b). In conclusion, several factors in­
fluence the performance of UHS, including injection and production
rates, the use of cushion gas, and relative permeability hysteresis, among
others. Understanding and accounting for these factors is crucial when
designing and implementing UHS projects. Moreover, further modelling
and experimental studies are essential to properly understand and
evaluate the effects of the various hydrodynamic parameters on UHS. A
summary of the modelling studies on the hydrodynamics of UHS is
presented in Table 6.
2.4. Emerging approaches
Approaches other than classical numerical simulations and
geochemical modelling have been used to model various aspects of UHS.
These approaches include Machine Learning (ML), Molecular Dynamics
Simulations (MD), Computational Fluid Dynamics (CFD), and Pore Scale
Modelling. In the following subsections, each approach will be briefly
introduced and discussed with the associated published modelling
studies.
2.4.1. Machine learning (ML)
ML has emerged as a vital tool in propelling the optimization and
modelling capabilities within UHS systems to new horizons. Its com­
petency in managing vast datasets and deciphering complex relation­
ships between variables positions it as a highly effective tool for
M. Saeed and P. Jadhawar Gas Science and Engineering 121 (2024) 205196

Table 6
A summary of studies on hydrodynamic modelling of UHS.
Reference Scope Storage formation type Approaches and Software Remarks
packages

Cai et al. (2022) Numerical simulation of UHS and gas Different scenarios including H2 PSFLOW (General Purpose Simulator was benchmarked against other
mixture storage storage in salt cavern, CH4- Subsurface Flow Simulator). simulators (TOUGH2/EWASG).
staurated aquifer, depleted gas Including: Results showed that capable of modelling
field with CO2 as cushion gas 1. Mass and energy non-isothermal, multiphase and
conservation equations for multicomponent flow in H2-water, H2–CH4-
gas flow. water, and H2–CH4–CO2-water systems at
2. Thermodynamics high pressure and temperature.
modelling.
Enigbokan et al. Numerical modelling to assess the Depleted gas reservoir Numerical simulation using They concluded that large-scale seasonal
(2021) feasibility of H2 storage in depleted gas Eclipse 300 storage can be achieved in the studied field.
reservoir
Heinemann et al. Numerical simulation to assess the role of Saline aquifer anticline CMG-GEM Cushion gas does not expand the storage
(2021b) cushion gas in saline aquifer anticline capacity; however, it can be used to exploit
the existing capacity efficiently hydrogen
storage in geological structures at greater
depths with higher reservoir permeabilities
requires lower amounts of cushion gas
Kanaani et al. Used numerical simulation to evaluate Depleted oil reservoirs CMG-GEM UHS without the use of cushion gas results
(2022) the effects of different types of cushion of up to 15.5% losses in the stored hydrogen
gases on the UHS in depleted oil Methane resulted in the higher hydrogen
reservoirs recovery compared to nitrogen and carbon
dioxide
Increased rates of hydrogen injection in
short intervals limits the effects of gravity
segregation due to improved sweeping of
reservoir fluids
Luboń and Numerical simulation to compare Deep aquifer anticline PetraSim TOUGH2 The studied structure can store up to 1
Tarkowski between the storage capacity of carbon million tons of carbon dioxide compared to
(2021) dioxide and hydrogen in deep aquifer 4000 tons hydrogen over a 31-year period.
anticline Altering the threshold capillary pressure of
the cap rock greatly influences the
modelling results
Mahdi et al. Numerical simulation of UHS in Heterogenous sandstone TOUGH2 coupled with Higher injection rates lead to increased
(2021a) heterogeneous sandstone reservoir reservoir EWASG hydrogen leakage in the absence of a sealing
caprock and increased hydrogen losses due
to undesirable gas fingering.
Pfeiffer and Bauer Numerical simulation of UHS to Underground natural gas Eclipse 300 Storage performance increases with the
(2015) investigate injection and production structure number of cycles and that shut-in periods
aspects during the initial filling period
Pfeiffer et al. Numerical simulation to evaluate Existing anticlinal structure in Eclipse 300 on average UHS can sustain a continuous
(2017) feasibility of UHS application the North German Basin power output of 245 MW for 1 week using
five storage wells.
Wang et al. Numerical simulation to investigate the Hypothetical reservoir CMG-GEM Dimensionless groups including aspect
(2022a) applicability of the scaling theory to ratio, gravity/viscous ratio and gravity/
hydrogen and carbon dioxide flow inside capillary ratio can be used to correlate the
a hypothetical reservoir. hydrogen and carbon dioxide flow between
different scales.
Heterogenous reservoirs may not be optimal
for UHS due to the drop in the purity of the
produced hydrogen, especially, at higher
injection and production rates.
Wang et al. Numerical simulation to investigate the Hypothetical reservoir CMG-GEM Due to CO2 solubility in brine, the volume of
(2022b) hydrogen and carbon dioxide flow inside injected carbon dioxide needs to be
a heterogenous reservoir and the effect of increased by approximately 30% to result in
carbon dioxide solubility in water on the the same displacement effect.
process Including carbon dioxide solubility in brine
in the model led to an increase in the
estimated hydrogen purity
Luboń and Numerical simulation of UHS to Deep aquifer structure in Poland PetraSim coupled with Water coning and excessive production is
Tarkowski investigate injection and production EWASG module the main hurdle for UHS in the aquifer
(2020) aspects hydrogen recovery is dependent on the
initial injection period, where a longer
initial period results in less hydrogen
recovery
Lysyy et al. Understand the effects of storing Norne hydrocarbon field Eclipse 100 UHS in the gas zone was the favoured option
(2021) hydrogen in three different zones, gas, oil (Norway) which resulted in a hydrogen recovery of
and water 87% compared to 77% and 49% hydrogen
recovery from oil and water zones,
respectively
Bai and Investigate the hydro-mechanical aspects Saline aquifer at the Powder 3D coupled hydro- A maximum of 75% hydrogen could be
Tahmasebi of UHS River Basin of Wyoming State mechanical model recovered by the end of the third cycle the
(2022) Peng-Robinson equation of studied structure is geomechanically safe
state
(continued on next page)

18
M. Saeed and P. Jadhawar Gas Science and Engineering 121 (2024) 205196

Table 6 (continued )
Reference Scope Storage formation type Approaches and Software Remarks
packages

the possibility of failure decreases during

the withdrawal period
Ebigbo et al. Investigated the effects of bioactivity on N/a Mathematical and numerical Possible competition between three
(2013) UHS modelling different microbial species: methanogens,
acetogens, and acetotrophs.
Pfeiffer et al. Numerical simulation of UHS in an Aquifer Eclipse 300 Storage performance increased with the
(2016) anticline structure increased number of storage cycles.
Hydrogen loss due to dissolution in connate
water was 1.75% of the initial hydrogen
storage over a 4-year period.
Hagemann et al. Assess hydrodynamic effects during UHS Aquifer Mathematical and Numerical Low injection rates result in gravitational
(2015) modelling utilising DuMuX forces becoming dominant leading to a
uniform water displacement.
UHS in stratified aquifers could limit the
risk of losing gas due to lateral spreading.
Feldmann et al. Investigate hydrodynamics and gas Depleted gas reservoir and Mathematical and Numerical Gravity override is more pronounced in
(2016) mixing during UHS aquifer modelling utilising DuMuX aquifers than in gas reservoirs.
Mechanical dispersion amplifies the gas
mixing process and is more pounced than
molecular diffusion.
Hagemann et al. Investigate biochemical and Depleted gas reservoir Numerical modelling The number of microorganisms present in
(2016) hydrodynamic behaviour during UHS. through DuMuX coupled the reservoir greatly influence the process of
with biochemical modelling hydrogen conversion into methane
Delshad et al. Investigated operational parameters Depleted oil reservoir and saline Numerical modelling using Cushion gas requirements in aquifer is
(2022b) effect on UHS aquifer CMG-GEM higher than the oil reservoir.
The presence of high permeability channel
in the oil reservoir requires additional wells
to contain hydrogen spreading
Delshad et al. Investigated operational parameters Depleted oil and gas reservoirs Numerical modelling using Results were sensitive to water/H2 relative
(2022a) effect on UHS and saline aquifer CMG-GEM permeability curves used in the aquifer case.
Hysteresis consideration resulted in higher
injectivity.
Molecular diffusion and solubility had
minimal impact on hydrogen saturation.
Abdellatif et al. Numerical simulation of hydrodynamics Depleted gas field CMG Simulation suite Lighter cushion gases results in higher
(2023a) during UHS hydrogen production.
Hydrogen diffusion is negligible.
Wang et al. Multi-scale simulation of UHS Depleted unconventional gas Monte Carlo Simulation, Hydrogen gas stored in unconventional gas
(2023b) reservoirs Numerical modelling using reservoirs maintains higher purity
In-house reservoir simulator compared to storage in conventional gas
MSCO2 reservoirs due to differential adsorption
effect of nanopores.
Most of the hydrogen is concentrated in the
hydraulically fractured region of reservoir.
Huang et al. Assess the viability of UHS Depleted gas reservoir TOUGH H2 loss due to diffusion is minimal (<1%).
(2023) Hydrogen recovery is significantly
improved with the use of N2 or CO2 as
cushion gases.
Jadhawar and Evaluated the effects of reservoir Deep aquifer Numerical modelling using Heterogeneity is an important factor in
Saeed (2023b) parameters and mechanisms on UHS CMG-GEM determining the performance of UHS.
Relative permeability hysteresis omission in
simulation may result in overestimating the
recoverable hydrogen gas.
Luboń and Evaluated the impact of the initial filling Deep aquifer PetraSim and TOUGH2 Extended initial filling period resulted in
Tarkowski period on the performance of UHS. improved operational storage capacity and
(2023) overall performance.
Jadhawar and Evaluated the impact of fluid-fluid and Deep aquifer CMG-GEM Hydrogen aqueous diffusion and solubility
Saeed (2023c) fluid-rock interactions during UHS. effects are negligible.
Jadhawar and The effect of operational parameters on Deep aquifer CMG-GEM The hydrogen recovery efficiency decreases
Saeed (2023a) UHS with the increase in operational rates,
longer and less frequent storage cycles,
presence of shut-in periods between
injection and production, and lower number
of operational wells.
Saeed and Studied the impact of cushion gas type on Aquifer CMG-GEM Utilising a cushion gas improves hydrogen
Jadhawar UHS. recovery efficiency.
(2023); Saeed Hydrogen recovery efficiency was directly
et al. (2023b) related to cushion gas density such that
lighter gases result in higher hydrogen
recovery.
CH4 resulted in higher hydrogen retrieval
compared to N2 and CO2.
Aftab et al. (2023) Assessment of hydrogen storage capacity Salt caverns AusH2 Results showed that 28,282 onshore salt
and viability caverns can be developed in Australia with
overall capacity of 14,697 PJ of hydrogen
energy.

19
M. Saeed and P. Jadhawar
modelling the intricate and interconnected processes intrinsic to UHS. A
notable application of ML algorithms is showcased in the work of Zhu
et al. (2022), who employed Wavelet Neural Networks (WNN) to model
hydrogen solubility in water under a variety of conditions, exemplifying
the adeptness of ML in accurately predicting complex behaviours in UHS
systems. Authors showed that modelling results predicted experimental
values with an R-squared value of 0.9999. Their results showed that
pressure is most impactful parameter on the hydrogen’s solubility in
brine followed by temperature and NaCl salinity. Ansari et al. (2022)
also utilised ML models to predict hydrogen solubility in water. They
found that a Radial Basis Function model (RBF) optimised using Cultural
Algorithm yielded the most accurate prediction of hydrogen solubility in
water under varying conditions. They also concluded that pressure was
the most impactful factor on hydrogen solubility which is similar the
observation of Zhu et al. (2022)).
ML has been used to model other fluid properties such as H2-water
IFT through different approaches. Wei Ng et al. (2022) reported that a
Multilayer Preceptor model optimised with Levenberg-Marquardt suc­
ceeded in predicting H2-water IFT with an R2 value of 0.9997 in a wide
range of salinity (0–4.95 mol/kg), pressure (0.5–45.2 MPa) and tem­
perature (298.03–448.35 K). Gbadamosi et al. (2023) showed that a
Gaussian Process Regression model outperforms other evaluated models
in predicting the IFT property in a H2-water system with a mean average
error of 0.022.
Moreover, ML stands as a robust mechanism for parameter optimi­
zation, which is a critical facet for bolstering the safety and efficiency of
UHS systems. By scrutinizing historical and real-time data, ML can assist
in pinpointing the optimal pressure and temperature conditions that
augment hydrogen storage capacity while mitigating the risk of leakage,
translating to more efficient operations and potentially diminished
operational costs.
One of the significant merits of ML lies in its capability to model and
integrate various intricate and interconnected processes like hydrody­
namic, fluid behaviour, geochemical, and geomechanical processes,
which are cardinal in comprehending the behaviour of hydrogen within
underground formations. For instance, in hydrodynamic processes, ML
can be employed to model the distribution and movement of hydrogen
within underground formations, aiding in predicting how hydrogen will
flow under various pressure, temperature, and wettability conditions,
which is crucial for the design and operation of UHS systems that aim for
efficient storage and retrieval of hydrogen. In this context, a study by Vo
Thanh et al. (2023), the use of machine learning algorithms, specifically
XGBoost, RF, LGRB, and Adaboost_DT, was investigated to predict
hydrogen wettability in underground storage sites using 513 data points.
The study found that the XGBoost algorithm was exceptionally profi­
cient, achieving an R2 value of 0.941 and RMSE and MAE values of
4.455 and 2.861 respectively. Substrate types were identified as the
most impactful variables influencing the predictions. Furthermore, the
study successfully predicted the hydrogen column height in a specific
basalt formation in Australia, showcasing the practical applicability and
efficiency of using machine learning in predicting hydrogen behaviour
and wettability in underground storage.
In tandem, the fluid behaviour of hydrogen within these formations
is a complex facet significantly influenced by the geomechanical and
geochemical conditions of the storage site. ML can help model and
predict this behaviour, thereby aiding in the design of storage caverns
and the optimization of storage and retrieval processes. Moreover, ML’s
utility extends to modelling geochemical interactions between hydrogen
and the surrounding rock formations. By analysing data on rock and
hydrogen interactions over time, ML algorithms can aid in forecasting
potential alterations in rock chemistry, which is imperative for ensuring
the long-term integrity and safety of UHS systems.
Simultaneously, understanding the geomechanical behaviour of rock
formations under the pressure and temperature conditions associated
with hydrogen storage is quintessential for ensuring the structural
integrity of the storage caverns. ML can be employed to model and
20
Gas Science and Engineering 121 (2024) 205196
analyse geomechanical data, aiding in the prediction of rock deforma­
tion and stress distributions within the underground formations, which
are critical variables for the safe and efficient operation of UHS systems.
These comprehensive insights, facilitated by ML, are instrumental in
advancing the design, operation, and maintenance of UHS systems, ul­
timately contributing to the evolution of safer and more efficient UHS
solutions.
2.4.2. Molecular dynamic simulations (MD)
MD simulations provide a detailed molecular-level analysis, which is
indispensable for understanding the interaction between hydrogen and
various materials in UHS systems. For instance, Ghasemi et al. (2022)
used molecular dynamic simulations to evaluate hydrogen diffusion in
three types of clay minerals (pyrophyllite, montmorillonite, and bei­
dellite) under varying clay charging behaviours. Their simulations were
performed using GROMACS. The modelling results showed that
hydrogen diffusion in clay minerals is significantly lower than that in
bulk water due to the confined environment of clays. Moreover, in
negatively charged clays, the increase in pore size up to 2 nm leads to an
increase in the hydrogen diffusion coefficient. However, above 2 nm, the
hydrogen diffusion coefficient remains constant. They also concluded
that the charging behaviour of clay minerals and interlayer cations af­
fects the hydrogen diffusion coefficient. Where the higher polarizability
of the O-sheet leads to the attraction of the water molecules and
consequently an increase in the diffusion coefficient. Their findings
suggest that the reservoir rock and caprock mineralogy need to be
properly assessed and evaluated to understand the potential hydrogen
diffusion during UHS. Moreover, MD simulations can model the solu­
bility of hydrogen within underground storage, providing insights into
how temperature, pressure, and material properties affect hydrogen
behaviour on a molecular level. Zhang et al. (2023) utilised a coupled
MD-ML approach to evaluate the solubility of hydrogen in brine under
varying salinities, pressures and temperatures. Their work indicated that
their approach was successful in reproducing the experimental results.
Moreover, they found that temperature has a non-linear effect on
hydrogen solubility in water.
MD simulations can be utilised to study the interaction between
hydrogen molecules and other substances like water or impurities
within the storage reservoir, which is crucial for understanding how
these interactions may affect hydrogen storage and retrieval. Further­
more, MD simulations have the capability to model the behaviour of
sealing materials under different temperatures and pressures, aiding in
the selection of materials that can maintain integrity under UHS oper­
ational conditions. This molecular-level understanding is essential for
designing UHS systems that can safely and efficiently store hydrogen
underground.
2.4.3. Computational fluid dynamics (CFD)
CFD models utilise numerical methods to analyse and solve problems
associated with fluid flows, allowing for the simulation of fluid dy­
namics, heat transfer, and chemical reactions within various systems.
Sáinz-García et al. (Sainz-Garcia et al., 2017a; Sainz-Garcia, 2017)
conducted simulations using COMSOL Multiphysics to study the
immiscible multiphase flow of water and a methane-hydrogen gas
mixture during methane-hydrogen underground storage in the Lower
Triassic of the Paris Basin. Their research indicated that both gas and
aquifer properties significantly influence the storage process. Moreover,
they developed a 3D multiphase numerical model to explore different
extraction well configurations, demonstrating that a maximum
hydrogen recovery of 78% could be achieved in the studied UHS case.
However, they also noted that hydrogen up-coning could pose a sub­
stantial challenge for UHS in saline aquifers without the use of cushion
gas. By employing tools like COMSOL for CFD simulations, gain crucial
insights can be gained into complex multiscale phenomena and opti­
mizing the design and operation of UHS facilities.
M. Saeed and P. Jadhawar
2.4.4. Pore network modelling
Pore Network Modelling is a simulation technique that abstracts the
intricate structure of porous materials into a simplified network of
interconnected void spaces, known as pores, and the constrictions be­
tween them, known as throats. This method facilitates the study of fluid
flow and transport phenomena within the porous media by capturing
essential features such as pore size, shape, and connectivity.
In 2021, Hashemi et al. (2021a) conducted a comprehensive study to
understand hydrogen transport characteristics in brine-saturated deep
porous rocks for UHS. They utilised pore-scale modelling and performed
a sensitivity analysis to quantify the effects of uncertain fluid and rock
properties on reservoir-scale functions, such as relative permeability and
capillary pressure. The study also explored cyclic hysteretic multiphase
flow, relevant to hydrogen-brine energy storage projects. Their results
showed that capillary pressure and relative permeability are sensitive to
contact angles. Moreover, clay content had a pronounced effect on the
endpoint values for relative permeability curves for drainage and
imbibition cycles. Wang et al. (2023c) investigated hydrogen transport
in sandstone porous media under different wetting conditions using
direct numerical simulation. The study found that increased hydrogen
wetting reduced the snap-off effect during the primary drainage process,
thereby facilitating greater hydrogen storage. However, during the
primary imbibition process, increased hydrogen wetting hindered the
extraction process, resulting in a recovery factor below 20%.
Bagheri et al. (2023) conducted a pore-scale investigation of
hydrogen-water flow dynamics in aquifers using computational fluid
dynamics at elevated pressure. The research identified that the optimal
rates for hydrogen injection and production differed, and a medium flow
rate could minimize both capillary and viscous fingering mechanisms,
thereby enhancing storability and recovery factor. The study also noted
that neglecting the local compression and expansion of hydrogen could
lead to misinterpretations of flow regimes and hydrogen storability.
Zhao et al. (2024) used a three-dimensional pore network model to
simulate hydrogen-brine two-phase flow in various porous media, such
as sandstone, carbonate, and sand packs. The study suggested that a
larger contact angle with low water affinity was preferable for UHS due
to its low hydrogen trapping rate. Machine learning methods, including
least square fitting and Support Vector Machine (SVM), were employed
to classify rocks based on their hydrogen trapping rates, providing in­
sights for site selection in UHS projects.
These studies highlight the importance of understanding hydrogen
transport and trapping mechanisms in porous media for effective UHS.
Both Wang et al. (2023c) and Bagheri et al. (2023) discuss the influence
of wetting conditions and flow rates on hydrogen storage and extraction,
while Hashemi et al. (2021a) and Zhao et al. (2024) emphasize the need
for a systematic understanding of fluid and rock properties on UHS. The
integration of pore scale modelling and machine learning methods by
Zhao et al. (2024) offers a new approach to predicting hydrogen trap­
ping rates, supplementing traditional pore network modelling tech­
niques. Such integrated approaches contribute to a comprehensive
understanding of factors influencing hydrogen storage in porous media
and provide valuable insights for UHS site selection and design.
2.4.5. Other approaches
Hassanpouryouzband et al. (2020) modelled thermodynamic and
transport properties of hydrogen containing mixtures. They used
GERG-2008 Equation of State (EoS) and SuperTRAPP model to predict
thermodynamic and transport properties of hydrogen mixed with
methane, carbon dioxide, nitrogen and natural gas (with a typical North
Sea natural gas composition). Their modelling results were validated
against experimental data from the literature and the model was found
to provide accurate predictions at pressures within the range 0.01–100
MPa, temperatures in the range 200–500 K and hydrogen mole fraction
in the range 10–90 Mole%. They developed a publicly accessible soft­
ware (H2Themobank) to calculate hydrogen gas mixture properties such
as gas density, thermal capacity, viscosity, entropy among other
21
Gas Science and Engineering 121 (2024) 205196
properties at varying mole fractions, temperatures and pressures. We
believe that their valuable effort will be of great benefit in advancing the
modelling of UHS and its associated processes. Researchers in HyStorPor
project developed an open-access tool to assess the hydrogen storage
capacities of both surface and subsurface facilities and formations in
varying conditions (Hassanpouryouzband et al., 2021).
Maraggi et al. (Ruiz Maraggi and Moscardelli, 2023) addressed the
lack of open-source tools for assessing H2 storage feasibility in caverns
by developing the GeoH2 Salt Storage and Cycling App. This web-based
thermodynamic simulator comprises modules for H2 physical properties,
volumetric capacity, production, injection, and cycling. Validation of
physical and volumetric modules with empirical data and comparison of
production and injection modules with an open-source simulenhance
the tool’s credibility. This App offers a means to assess technical aspects
of H2 storage, injection, withdrawal, and cycling in salt caverns. A
summary of various studies that utilised modelling approaches other
than the geochemical, geomechanical, and hydrodynamic approaches
are shown in Table 7.
3. Current status of modelling studies
In reviewing the current landscape of UHS modelling studies as
shown in Fig. 12a, it is evident that research efforts have been unevenly
distributed across different aspects of the discipline. The analysis of
recent modelling studies indicates a pronounced emphasis on hydro­
dynamic aspects, as they receive the most attention. This focus points to
the critical importance placed on understanding the subsurface flow and
distribution of hydrogen, a fundamental component in assessing the
viability and safety of UHS.
Conversely, geochemical considerations, have been less explored in
modelling studies compared to hydrodynamics. These studies are
essential for comprehending the chemical interactions between stored
hydrogen and the geological media, which have implications for both
the integrity of the storage formation and the quality of the stored
hydrogen. Geomechanical factors, which assess the mechanical impacts
of hydrogen storage on the surrounding geology, have been the subject
of relatively few studies, suggesting a potential gap in current research
efforts.
The selection of tools for modelling these storage systems (see
Fig. 12b) reveals a preference for the Computer Modelling Group’s
(CMG) suite of simulation software, which has been the most widely
adopted among the tools assessed. Its extensive use implies that CMG’s
capabilities are well-aligned with the current focus of the field.
Furthermore, the application of Machine Learning (ML) approaches in a
substantial number of studies reflects an innovative trend in the field,
harnessing the power of data-driven techniques to enhance model ac­
curacy and efficiency. Other tools, including PHREEQC for geochemical
calculations and the various other software packages represented in the
present modelling studies, show a diversity in the methodologies being
employed in current research. This indicates a field that values and re­
quires the application of multiple approaches to understand the multi­
faceted aspects encountered during UHS.
4. Challenges and future research
The present investigation into UHS modelling has revealed several
issues and challenges that remain unresolved and are summarised
below:
• One of the significant challenges in UHS modelling is the pronounced
lack of expansive and diverse datasets. Currently, the reliance on
limited field trials and laboratory studies for model validation fails to
provide the scope and specificity needed for confident application in
diverse geological settings. This insufficiency leads to a significant
gap between model predictions and their practical reliability.
Therefore, a concerted effort is necessary to compile comprehensive
M. Saeed and P. Jadhawar
Table 7
A summary of studies that utilised alternative modelling of UHS.
Reference
Gajda and Lutyński (2022)
Zhu et al. (2022)
Ghasemi et al. (2022)
Hassanpouryouzband et al.
(2020)
Ruiz Maraggi and
Moscardelli (2023)
Kanaani et al. (2023)
Ansari et al. (2022)
Ng et al. (2022)
Gbadamosi et al. (2023)
Vo Thanh et al. (2023)
Zhang et al. (2023)
(Sainz-Garcia et al., 2017a;
Sainz-Garcia, 2017)
Hashemi et al. (2021a)
Wang et al. (2023c)
Bagheri et al. (2023)
Zhao et al. (2024)
Hassanpouryouzband et al.
(2021)
Scope
Modelling permeability and
hydrogen loss through Polymer
Sealing Liners during UHS
Predict hydrogen solubility in
water as a function of pressure,
temperature and NaCl salinity
Evaluate hydrogen diffusion
Modelling thermodynamic and
transport properties of
hydrogen containing mixtures
Modelling of UHS in salt
caverns
Underground hydrogen gas
storage and carbon dioxide
storage optimization
Modelling hydrogen solubility
in water
Modelled H2-water IFT
Modelled H2-water IFT
Modelled hydrogen wettability
in UHS sites
Evaluated solubility of
hydrogen in brine
Studied immiscible multiphase
flow of water and a methane-
hydrogen gas mixture
Evaluated impacts of uncertain
rock-fluid properties
Hydrogen transport in different
wetting conditions
Hydrogen flow behaviour
Hydrogen-brine two-phase
flow in various porous media
Hydrogen storage capacity
Storage formation type/
Conditions
Lined Rock Cavern (LRC)
Underground sodium chloride
solutions
Three types of clay minerals
(pyrophyllite, montmorillonite,
and beidellite)
– GERG-2008 Equation of State
Salt caverns
Synthetic reservoir model
Varying salinity, pressure and
temperature conditions
Salinity (0–4.95 mol/kg),
pressure (0.5–45.2 MPa) and
temperature (298.03–448.35
K)
Pressure (2.76–34.47 MPa) and
temperature (298.15–423.15
K)
Various sites
Varying salinities, pressures
and temperatures
Aquifer in Lower Triassic of the
Paris Basin
Deep aquifer
Aquifer
Aquifer
Sandstone, carbonate, and sand
pack
Surface facilities and
subsurface formations
22
Gas Science and Engineering 121 (2024) 205196
Approaches and Software Remarks
packages
– Model was successful in modelling
permeability and hydrogen loss within
studied conditions
ML Used WNN to model solubility.
Pressure is the most impactful parameter on
the hydrogen’s solubility in brine followed
by temperature and NaCl salinity
MD (GROMACS) hydrogen diffusion in clay minerals is
significantly lower than that in bulk water
due to the confined environment of clays
reservoir rock and caprock mineralogy needs
to be properly assessed and evaluated to
understand the potential hydrogen diffusion
during UHS
Modelling results were validated against
(EoS) and supertrapp model experimental data from the literature.
(H2Themobank) Developed a publicly accessible software to
calculate hydrogen gas mixtures properties
Thermodynamical simulator The developed tool used empirical data for
(GeoH2 Salt Storage and validation. Serves as a means for evaluating
Cycling App) the physical attributes of hydrogen during
UHS, injection, and withdrawal.
ML Optimization of UHS, carbon dioxide
sequestration and net present value as
objective functions show conflict between
the three functions.
ML Radial Basis Function model (RBF) optimised
using Cultural Algorithm yielded the most
accurate prediction of hydrogen solubility in
water.
Pressure was the most impactful factor on
hydrogen solubility
ML Multilayer Preceptor model optimised with
Levenberg-Marquardt succeeded in
predicting H2-water IFT with an R2 value of
0.9997.
ML Utilised a Gaussian Process Regression model
to predict IFT with a mean average error of
0.022.
ML Xgboost model successfully predicted the
hydrogen column height in a specific basalt
formation in Australia.
Substrate types were identified as the most
impactful variables influencing the
predictions.
MD-ML Their approach was successful in
reproducing the experimental results.
Temperature has a non-linear effect on
hydrogen solubility in water.
CFD (COMSOL) Maximum hydrogen recovery of 78% could
be achieved in the studied UHS case.
Hydrogen up-coning could pose a substantial
challenge for UHS in saline aquifers without
the use of cushion gas.
Pore Network Model Contact angle (wettability) effect on relative
permeability was observed.
Clay content affected the relative
permeability curves.
Pore network model Higher hydrogen wetting reduced the snap-
off effect during the primary drainage
process and increased storage capacity but
decreased the recovery efficiency.
Pore network model, CFD Medium flow rate could minimize both
capillary and viscous fingering mechanisms.
Important to consider local hydrogen
compressibility to properly identify flow
regime.
Pore network model, ML Larger contact angle with low water affinity
was preferable for UHS due to its low
hydrogen trapping rate.
Nobel-Abel Equation of State, This open-source tool allows for quick
C# code in visual studio calculation of hydrogen storage capacity in
(H2capES) both surface and subsurface facilities and
pore paces.
M. Saeed and P. Jadhawar
Fig. 12. Distribution of modelling studies based on (a) the focus of study and
(b) utilised modelling tool.
datasets that reflect hydrogen behaviour under a broader spectrum
of geological variables and operational conditions. This is imperative
for developing more accurate and reliable models capable of guiding
the implementation of UHS projects. Current open-source databanks
and UHS software tools such as NIST Chemistry WebBook, the
Dortmund Data Bank, H2CapES, H2Thermobank, and GeoH2 Salt
Storage and Cycling App, can provide valuable data for UHS
modelling. Developing similar open-source databanks and tools for
the various hydrodynamics, geochemical and geomechanical
experimental data and field data will be of value for researchers and
engineers in the UHS field.
• The process of upscaling from precise laboratory data to the unpre­
dictable variability of field-scale conditions necessitates the devel­
opment of robust methodologies. While laboratory models offer
high-resolution insights within controlled settings, they do not
inherently capture the complex heterogeneity found in natural sub­
surface environments. To bridge this gap, it is essential to formulate
methodologies that can adapt and scale laboratory findings to align
with field data. This upscaling is not merely a question of increasing
data volume but requires a sophisticated integration of multi-scale
processes. Effective methods must be established to ensure that
laboratory models accurately reflect and can be validated against the
wider range of conditions encountered during field trials. Such
methodologies would enable a reliable comparison between highly
controlled laboratory results and the diverse datasets derived from
field operations, leading to models that can be trusted for real-world
application and analysis.
• The accurate representation of UHS systems through modelling ne­
cessitates intensive computational resources, particularly when
integrating diverse aspects such as hydrodynamics, geochemistry,
23
Gas Science and Engineering 121 (2024) 205196
and geomechanics. High-performance computing (HPC) platforms
become a prerequisite as the models strive to capture an array of
complex interactions over disparate scales, from the molecular level
to the full extent of the storage reservoir. Moreover, there is a
pressing need for the development of novel computational strategies,
potentially through the application of parallel computing, machine
learning and artificial intelligence, to enhance the efficiency of
simulations and make the most effective use of available computa­
tional resources.
• A notable gap in the current body of research within UHS is the lack
of modelling studies that address the geological uncertainties
inherent in subsurface environments. Addressing the underexplored
area of geological uncertainty is paramount for enhancing UHS
modelling. The complexity of subsurface geology, with its unpre­
dictable facies and fault structures, introduces significant variability
in predicting porosity, permeability, and containment integrity.
Although advanced geostatistical and geophysical methods exist
within oil reservoir simulation, their application to UHS remains
sparse. Adapting these established methodologies could provide a
shortcut to more accurate geological models for UHS. Emphasizing
this research gap is essential, as it underpins the accurate assessment
of storage capacities and operational safety. By drawing on oil and
gas industry experiences, we can potentially fast-track the develop­
ment of UHS models that effectively navigate geological
uncertainties.
This review focused on the technical modelling of UHS; however,
another important aspect of modelling UHS is the economic modelling
of the process. This type of modelling should involve the full value chain
of UHS to allow for a proper comparison between UHS and alternative
large scale energy storage systems. There exist some economic studies
on this topic, however, they are beyond the scope of this study.
5. Conclusion
The present study conducted an extensive investigation into the
current approaches and studies on UHS modelling. The following con­
clusions summarise the most important findings of this review:
• Modelling UHS is a complex process that requires an understanding
and integration of various multi-scale phenomena and aspects,
including hydrodynamics, geochemistry, and geomechanics.
• Geochemical studies on UHS predominantly utilised PHREEQC or
Geochemists Workbench to describe geochemical reactions. The
amount of hydrogen gas loss as a result of abiotic and biotic reactions
can range from negligible losses up to more than 80%, depending on
the adopted geochemical reactions and kinetic reaction rates, simu­
lated conditions, rock mineralogy, brine salinity, and most impor­
tantly, the adoption of hydrogen redox reactions.
• The 3D coupled hydro-mechanical models revealed that rock failure
during the withdrawal period in UHS in aquifers decreases compared
to the injection period due to a decrease in pore pressure and an
increase in effective stress. Additionally, the geomechanical changes
during UHS in salt caverns depend on the number of storage cycles,
temperature, presence of impurities in the salt cavern, as well as
injection and production rates.
• The most widely used software programs for modelling hydrogen
hydrodynamics are CMG’s simulation suite, TOUGH, DuMux, and
Eclipse. These simulators use mathematical and numerical models
that adopt fundamental fluid flow in porous media equations such as
conservation of mass, Darcy’s law, Fick’s law for diffusion modelling,
and Peng-Robinson EOS to model hydrogen gas properties. Hydro­
dynamics simulation studies revealed that the shape and rock
properties of the underground storage location play a critical role in
determining the hydrogen recovery efficiency. Higher injection and
production rates result in viscous forces becoming more dominant
M. Saeed and P. Jadhawar
than gravity forces, leading to higher hydrogen gas losses. The uti­
lization of cushion gas was also shown to be preferred, as it improves
hydrogen recovery efficiency. It is worth noting that the current
simulation modelling approaches are adapted from oil and gas
reservoir simulation approaches, as well as underground gas storage,
such as natural gas and carbon dioxide. The emerging approaches in
modelling UHS systems extend beyond classical numerical simula­
tions and geochemical modelling, encompassing Machine Learning
(ML), Molecular Dynamics Simulations (MD), Computational Fluid
Dynamics (CFD), Pore Scale Modelling, and other novel techniques.
ML has proven invaluable for deciphering complex relationships and
optimizing parameters in UHS systems, while MD simulations pro­
vide a molecular-level understanding of hydrogen interactions
crucial for system design. CFD models offer insights into fluid dy­
namics and heat transfer, enhancing the optimization of UHS facil­
ities. Pore Network Modelling, on the other hand, abstracts the
porous structure of materials to study fluid flow, which is crucial for
understanding hydrogen transport in porous rocks.
• Alternative approaches such as the development of open-source tools
and software for assessing thermodynamic and transport properties
of hydrogen mixtures and storage capacities significantly contribute
to the progression and optimization of UHS systems. Collectively,
these approaches and tools not only advance the understanding of
hydrogen behaviour and interactions in subsurface environments but
also play a pivotal role in realizing efficient and safe UHS solutions.
The present study provides crucial insights into the complexities and
challenges involved in modelling UHS. The findings of this study can
guide future research efforts to improve the accuracy and reliability of
the modelling approaches and studies, which is critical for the efficient
and safe implementation of UHS as a clean energy solution.
CRediT authorship contribution statement
Motaz Saeed: Conceptualization, Data curation, Formal analysis,
Methodology, Visualization, Writing - original draft, Writing - review &
editing. Prashant Jadhawar: Conceptualization, Data curation, Formal
analysis, Funding acquisition, Investigation, Methodology, Project
administration, Resources, Supervision, Validation, Visualization,
Writing - original draft, Writing - review & editing.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Data availability
No data was used for the research described in the article.
Acknowledgement
The authors gratefully acknowledge the funding support by the Net
Zero Technology Centre (NZTC), UK and the industrial sponsors to
accomplish this work under the Hydrogen Innovation Grant scheme.
References
Abdellatif, M., Hashemi, M., Azizmohammadi, S., 2023. Large-scale underground
hydrogen storage: integrated modeling of reservoir-wellbore system. Int. J.
Hydrogen Energy. https://doi.org/10.1016/J.IJHYDENE.2023.01.227.
Abe, J.O., Popoola, A.P.I., Ajenifuja, E., Popoola, O.M., 2019. Hydrogen energy,
economy and storage: review and recommendation. Int. J. Hydrogen Energy 44,
15072–15086. https://doi.org/10.1016/J.IJHYDENE.2019.04.068.
Aftab, A., Hassanpouryouzband, A., Xie, Q., Machuca, L.L., Sarmadivaleh, M., 2022.
Toward a fundamental understanding of geological hydrogen storage. Ind. Eng.
Chem. Res. 61, 3233–3253. https://doi.org/10.1021/ACS.IECR.1C04380.
24
Gas Science and Engineering 121 (2024) 205196
Aftab, A., Hassanpouryouzband, A., Naderi, H., Xie, Q., Sarmadivaleh, M., 2023.
Quantifying onshore salt deposits and their potential for hydrogen energy storage in
Australia. J. Energy Storage 65, 107252. https://doi.org/10.1016/J.
EST.2023.107252.
Ahluwalia, R.K., Wang, X., Rousseau, A., 2005. Fuel economy of hybrid fuel-cell vehicles.
J. Power Sources 152, 233–244. https://doi.org/10.1016/j.jpowsour.2005.01.052.
Ahmed, T., 2010. Reservoir Engineering Handbook. Elsevier Inc.
Al-Awad, M.N.J., 2002. Simple correlation to evaluate Mohr-Coulomb failure criterion
using Uniaxial compressive strength. Journal of King Saud University - Engineering
Sciences 14, 137–144. https://doi.org/10.1016/S1018-3639(18)30749-9.
Ali, M., Yekeen, N., Pal, N., Keshavarz, A., Iglauer, S., Hoteit, H., 2021. Influence of
pressure, temperature and organic surface concentration on hydrogen wettability of
caprock; implications for hydrogen geo-storage. Energy Rep. 7, 5988–5996. https://
doi.org/10.1016/J.EGYR.2021.09.016.
Amid, A., Mignard, D., Wilkinson, M., 2016. Seasonal storage of hydrogen in a depleted
natural gas reservoir. Int. J. Hydrogen Energy 41, 5549–5558. https://doi.org/
10.1016/j.ijhydene.2016.02.036.
Aminudin, M.A., Kamarudin, S.K., Lim, B.H., Majilan, E.H., Masdar, M.S., Shaari, N.,
2023a. An overview: current progress on hydrogen fuel cell vehicles. Int. J.
Hydrogen Energy 48, 4371–4388. https://doi.org/10.1016/J.
IJHYDENE.2022.10.156.
Aminudin, M.A., Kamarudin, S.K., Lim, B.H., Majilan, E.H., Masdar, M.S., Shaari, N.,
2023b. An overview: current progress on hydrogen fuel cell vehicles. Int. J.
Hydrogen Energy 48, 4371–4388. https://doi.org/10.1016/J.
IJHYDENE.2022.10.156.
Ansari, S., Safaei-Farouji, M., Atashrouz, S., Abedi, A., Hemmati-Sarapardeh, A.,
Mohaddespour, A., 2022. Prediction of hydrogen solubility in aqueous solutions:
comparison of equations of state and advanced machine learning-metaheuristic
approaches. Int. J. Hydrogen Energy 47, 37724–37741. https://doi.org/10.1016/J.
IJHYDENE.2022.08.288.
Bagheri, M., Mahani, H., Ayatollahi, S., Zivar, D., 2023. Direct pore-scale simulation of
the effect of capillary number and gas compressibility on cyclic underground
hydrogen storage & production in heterogeneous aquifers. Adv. Water Resour. 181,
104547 https://doi.org/10.1016/J.ADVWATRES.2023.104547.
Bai, T., Tahmasebi, P., 2022. Coupled hydro-mechanical analysis of seasonal
underground hydrogen storage in a saline aquifer. J. Energy Storage 50, 104308.
Bernardez, L.A., De Lima, L.R.P.A., De Jesus, E.B., Ramos, C.L.S., Almeida, P.F., 2013.
A kinetic study on bacterial sulfate reduction. Bioproc. Biosyst. Eng. 36, 1861–1869.
https://doi.org/10.1007/S00449-013-0960-0/TABLES/3.
Bethke, C.M., 1996. Geochemical reaction modeling: concepts and applications.
Geochemical Reaction Modeling. https://doi.org/10.1093/OSO/
9780195094756.001.0001.
Bethke, C.M., Fb, Sm, 2021. The Geochemist’s Workbench® Release 16 (Five Volumes).
Aqueous Solutions LLC, Champaign, Illinois.
Blyton, C.A.J., Gala, D.P., Sharma, M.M., 2015. A comprehensive study of Proppant
transport in a hydraulic fracture. Proceedings - SPE Annual Technical Conference
and Exhibition 2015. https://doi.org/10.2118/174973-MS. January:3391–411.
Bo, Z., Zeng, L., Chen, Y., Xie, Q., 2021. Geochemical reactions-induced hydrogen loss
during underground hydrogen storage in sandstone reservoirs. Int. J. Hydrogen
Energy 46, 19998–20009. https://doi.org/10.1016/j.ijhydene.2021.03.116.
Bo, Z., Boon, M., Hajibeygi, H., Hurter, S., 2023. Impact of experimentally measured
relative permeability hysteresis on reservoir-scale performance of underground
hydrogen storage (UHS). Int. J. Hydrogen Energy. https://doi.org/10.1016/j.
ijhydene.2022.12.270.
Boon, M., Hajibeygi, H., 2022. Experimental characterization of H2/water multiphase
flow in heterogeneous sandstone rock at the core scale relevant for underground
hydrogen storage (UHS). Sci. Rep. 12, 1. https://doi.org/10.1038/s41598-022-
18759-8, 2022;12:1–12.
Böttcher, N., Görke, U.-J., Kolditz, O., Nagel, T., 2017. Thermo-mechanical investigation
of salt caverns for short-term hydrogen storage. Environ. Earth Sci. 76, 1–13.
Brady, N.C., Weil, R.R., 2016. Nature and Properties of Soils, The fifteenth ed. Pearson
Education, p. 1104.
Cai, Z., Zhang, K., Guo, C., 2022. Development of a novel simulator for modelling
underground hydrogen and gas mixture storage. Int. J. Hydrogen Energy 47,
8929–8942. https://doi.org/10.1016/J.IJHYDENE.2021.12.224.
Cavallaro, A.N., Gracia Martinez, M.E., Ostera, H., Panarello, H., Cordero, R.R., 2005.
Oilfield reservoir souring during waterflooding: a case study with low Sulphate
concentration in formation and injection waters. SPE International Symposium on
Oilfield Chemistry Proceedings. https://doi.org/10.2118/92959-MS.
Chabab, S., Théveneau, P., Coquelet, C., Corvisier, J., Paricaud, P., 2020. Measurements
and predictive models of high-pressure H2 solubility in brine (H2O+NaCl) for
underground hydrogen storage application. Int. J. Hydrogen Energy 45,
32206–32220. https://doi.org/10.1016/J.IJHYDENE.2020.08.192.
Charlet, L., Alt-Epping, P., Wersin, P., Gilbert, B., 2017. Diffusive transport and reaction
in clay rocks: a storage (nuclear waste, CO2, H2), energy (shale gas) and water
quality issue. Adv. Water Resour. 106, 39–59. https://doi.org/10.1016/J.
ADVWATRES.2017.03.019.
Chen, Z., 2007. Reservoir Simulation: Mathematical Techniques in Oil Recovery. SIAM.
Coarita-Tintaya, E.-D., Golfier, F., Grgic, D., Souley, M., Cheng, L., 2023.
Hydromechanical modelling of salt caverns subjected to cyclic hydrogen injection
and withdrawal. Comput. Geotech. 162, 105690 https://doi.org/10.1016/J.
COMPGEO.2023.105690.
Computer Modelling Group, 2022. CMG User Manual. Computer Modelling Group,
Calgary, Canada.
Dake, L.P., 1978. Fundamentals of Reservoir Engineering. Elsevier.
M. Saeed and P. Jadhawar
Dawood, F., Anda, M., Shafiullah, G.M., 2020. Hydrogen production for energy: an
overview. Int. J. Hydrogen Energy 45, 3847–3869. https://doi.org/10.1016/J.
IJHYDENE.2019.12.059.
Delshad, M., Alhotan, M., Fernandes, B.R.B., Umurzakov, Y., Sepehrnoori, K., 2022a.
Pros and Cons of saline aquifers against depleted hydrocarbon reservoirs for
hydrogen energy storage. Proceedings - SPE Annual Technical Conference and
Exhibition 2022. https://doi.org/10.2118/210351-MS. October:3–5.
Delshad, M., Umurzakov, Y., Sepehrnoori, K., Eichhubl, P., Batista Fernandes, B.R.,
2022b. Hydrogen storage assessment in depleted oil reservoir and saline aquifer.
Energies 15, 8132. https://doi.org/10.3390/EN15218132.
Dickinson, E.J.F., Ekström, H., Fontes, E., 2014. COMSOL Multiphysics®: finite element
software for electrochemical analysis. A mini-review. Electrochem. Commun. 40,
71–74. https://doi.org/10.1016/J.ELECOM.2013.12.020.
Ebigbo, A., Golfier, F., Quintard, M., 2013. A coupled, pore-scale model for
methanogenic microbial activity in underground hydrogen storage. Adv. Water
Resour. 61, 74–85. https://doi.org/10.1016/J.ADVWATRES.2013.09.004.
Enigbokan, T., Heinemann, N., Li, J., 2021. Application of CO2 geologic storage
experience to underground hydrogen storage reservoirs. SSRN Electron. J. https://
doi.org/10.2139/SSRN.3821362.
Fang, Y., Hou, Z., Yue, Y., Chen, Q., Liu, J., 2022. Numerical Study of Hydrogen Storage
Cavern in Thin-Bedded Rock Salt, Anning of China. CRC Press, pp. 652–661.
Feldmann, F., Hagemann, B., Ganzer, L., Panfilov, M., 2016. Numerical simulation of
hydrodynamic and gas mixing processes in underground hydrogen storages.
Environ. Earth Sci. 75 https://doi.org/10.1007/S12665-016-5948-Z.
Flemisch, B., Darcis, M., Erbertseder, K., Faigle, B., Lauser, A., Mosthaf, K., et al., 2011.
DuMux: DUNE for multi-{phase, component, scale, physics, …} flow and transport in
porous media. Adv. Water Resour. 34, 1102–1112. https://doi.org/10.1016/J.
ADVWATRES.2011.03.007.
Flesch, S., Pudlo, D., Albrecht, D., Jacob, A., Enzmann, F., 2018. Hydrogen underground
storage—Petrographic and petrophysical variations in reservoir sandstones from
laboratory experiments under simulated reservoir conditions. Int. J. Hydrogen
Energy 43, 20822–20835. https://doi.org/10.1016/J.IJHYDENE.2018.09.112.
Gaines, G.L., Thomas, H.C., 1953. Adsorption studies on clay minerals. II. A formulation
of the thermodynamics of exchange Adsorption. J. Chem. Phys. 21, 714–718.
https://doi.org/10.1063/1.1698996.
Gajda, D., Lutyński, M., 2022. Permeability modeling and estimation of hydrogen loss
through Polymer sealing Liners in underground hydrogen storage. Energies 15,
2663. https://doi.org/10.3390/EN15072663, 2022;15:2663.
Gallo, A.B., Simões-Moreira, J.R., Costa, H.K.M., Santos, M.M., Moutinho dos Santos, E.,
2016. Energy storage in the energy transition context: a technology review. Renew.
Sustain. Energy Rev. 65, 800–822. https://doi.org/10.1016/J.RSER.2016.07.028.
Gbadamosi, A., Adamu, H., Usman, J., Usman, A.G., Jibril, M.M., Salami, B.A., et al.,
2023. New-generation machine learning models as prediction tools for modeling
interfacial tension of hydrogen-brine system. Int. J. Hydrogen Energy. https://doi.
org/10.1016/J.IJHYDENE.2023.09.170.
Ghasemi, M., Omrani, S., Mahmoodpour, S., Zhou, T., 2022. Molecular dynamics
simulation of hydrogen diffusion in water-saturated clay minerals; implications for
Underground Hydrogen Storage (UHS). Int. J. Hydrogen Energy 47, 24871–24885.
https://doi.org/10.1016/J.IJHYDENE.2022.05.246.
Gram, T.B., Ditlevsen, F.P., Mosegaard, K., Fabricius, I.L., 2021. Water-flooding and
consolidation of reservoir chalk – effect on porosity and Biot’s coefficient. Geophys.
Prospect. 69, 495–513. https://doi.org/10.1111/1365-2478.13047/CITE/BIBTEX.
Gregory, S.P., Barnett, M.J., Field, L.P., Milodowski, A.E., 2019. Subsurface microbial
hydrogen cycling: natural occurrence and implications for industry. Microorganisms
7, 53. https://doi.org/10.3390/MICROORGANISMS7020053, 2019;7:53.
Hagemann, B., Rasoulzadeh, M., Panfilov, M., Ganzer, L., Reitenbach, V., 2015.
Mathematical modeling of unstable transport in underground hydrogen storage.
Environ. Earth Sci. 73, 6891–6898. https://doi.org/10.1007/S12665-015-4414-7.
Hagemann, B., Rasoulzadeh, M., Panfilov, M., Ganzer, L., Reitenbach, V., 2016.
Hydrogenization of underground storage of natural gas: impact of hydrogen on the
hydrodynamic and bio-chemical behavior. Comput. Geosci. 20, 595–606. https://
doi.org/10.1007/S10596-015-9515-6.
Hashemi, L., Blunt, M., Hajibeygi, H., 2021a. Pore-scale modelling and sensitivity
analyses of hydrogen-brine multiphase flow in geological porous media. Sci. Rep. 11
https://doi.org/10.1038/S41598-021-87490-7.
Hashemi, L., Glerum, W., Farajzadeh, R., Hajibeygi, H., 2021b. Contact angle
measurement for hydrogen/brine/sandstone system using captive-bubble method
relevant for underground hydrogen storage. Adv. Water Resour. 154 https://doi.
org/10.1016/J.ADVWATRES.2021.103964.
Hassannayebi, N., Azizmohammadi, S., De Lucia, M., Ott, H., 2019. Underground
hydrogen storage: application of geochemical modelling in a case study in the
Molasse Basin, Upper Austria. Environ. Earth Sci. 78, 177. https://doi.org/10.1007/
s12665-019-8184-5.
Hassanpouryouzband, A., Joonaki, E., Edlmann, K., Heinemann, N., Yang, J., 2020.
Thermodynamic and transport properties of hydrogen containing streams. Sci. Data
7, 1–14. https://doi.org/10.1038/s41597-020-0568-6, 2020 7:1.
Hassanpouryouzband, A., Joonaki, E., Edlmann, K., Haszeldine, R.S., 2021. Offshore
geological storage of hydrogen: is this our best option to achieve net-zero? ACS
Energy Lett. 6, 2181–2186. https://doi.org/10.1021/ACSENERGYLETT.1C00845.
Hassanpouryouzband, A., Adie, K., Cowen, T., Thaysen, E.M., Heinemann, N., Butler, I.
B., et al., 2022. Geological hydrogen storage: geochemical reactivity of hydrogen
with sandstone reservoirs. ACS Energy Lett. 7, 2203–2210. https://doi.org/10.1021/
ACSENERGYLETT.2C01024/ASSET/IMAGES/LARGE/NZ2C01024_0004.JPEG.
Hawez, H.K., Sanaee, R., Faisal, N.H., 2021. A critical review on coupled geomechanics
and fluid flow in naturally fractured reservoirs. J. Nat. Gas Sci. Eng. 95, 104150
https://doi.org/10.1016/J.JNGSE.2021.104150.
25
Gas Science and Engineering 121 (2024) 205196
Heinemann, N., Alcalde, J., Miocic, J.M., Hangx, S.J.T., Kallmeyer, J., Ostertag-
Henning, C., et al., 2021a. Enabling large-scale hydrogen storage in porous media-
the scientific challenges. Energy Environ. Sci. 14, 853–864. https://doi.org/
10.1039/D0EE03536J.
Heinemann, N., Scafidi, J., Pickup, G., Thaysen, E.M., Hassanpouryouzband, A.,
Wilkinson, M., et al., 2021b. Hydrogen storage in saline aquifers: the role of cushion
gas for injection and production. Int. J. Hydrogen Energy 46, 39284–39296. https://
doi.org/10.1016/J.IJHYDENE.2021.09.174.
Hematpur, H., Abdollahi, R., Rostami, S., Haghighi, M., Blunt, M.J., 2023. Review of
underground hydrogen storage: concepts and challenges. Advances in Geo-Energy
Research 7, 111–131. https://doi.org/10.46690/AGER.2023.02.05.
Hemme, C., van Berk, W., 2018. Hydrogeochemical modeling to identify potential risks
of underground hydrogen storage in depleted gas fields. Appl. Sci. 8, 2282. https://
doi.org/10.3390/app8112282.
Hermesmann, M., Müller, T.E., 2022. Green, Turquoise, blue, or grey? Environmentally
friendly hydrogen production in Transforming energy systems. Prog. Energy
Combust. Sci. 90, 100996 https://doi.org/10.1016/J.PECS.2022.100996.
Hosseini, M., Fahimpour, J., Ali, M., Keshavarz, A., Iglauer, S., 2022. H2− brine
interfacial tension as a function of salinity, temperature, and pressure; implications
for hydrogen geo-storage. J. Pet. Sci. Eng. 213, 110441 https://doi.org/10.1016/J.
PETROL.2022.110441.
Huang, T., Moridis, G.J., Blasingame, T.A., Abdulkader, A.M., Yan, B., 2023.
Compositional reservoir simulation of underground hydrogen storage in depleted gas
reservoirs. Int. J. Hydrogen Energy. https://doi.org/10.1016/J.
IJHYDENE.2023.05.355.
International Energy Agency, 2022. Global Hydrogen Review 2022 – Analysis - IEA.
Jadhawar, P., Saeed, M., 2023a. Optimizing the operational efficiency of the
underground hydrogen storage scheme in a deep North Sea aquifer through
compositional simulations. J. Energy Storage 73, 108832. https://doi.org/10.1016/
j.est.2023.108832.
Jadhawar, P., Saeed, M., 2023b. Mechanistic evaluation of the reservoir engineering
performance for the underground hydrogen storage in a deep North Sea aquifer. Int.
J. Hydrogen Energy. https://doi.org/10.1016/J.IJHYDENE.2023.07.272.
Jadhawar, P., Saeed, M., 2023c. Fluid-Fluid and Rock-Fluid Parametric Investigation of
the Underground Storage of Hydrogen in a Deep North Sea Aquifer. https://doi.org/
10.2118/215598-MS.
Joseferd, R., 2015. IMPORTANCE OF POROSITY-PERMEABILITY RELATIONSHIP IN
SANDSTONE: PETROPHYSICAL PROPERTIES. Universiti Teknologi PETRONAS.
Kanaani, M., Sedaee, B., 2022. Impact of Dilation and Irreversible Compaction on
underground hydrogen storage in depleted hydrocarbon reservoirs. Energy Fuel. 36,
13506–13517. https://doi.org/10.1021/ACS.ENERGYFUELS.2C02150/ASSET/
IMAGES/LARGE/EF2C02150_0017.JPEG.
Kanaani, M., Sedaee, B., Asadian-Pakfar, M., 2022. Role of cushion gas on underground
hydrogen storage in depleted oil reservoirs. J. Energy Storage 45. https://doi.org/
10.1016/J.EST.2021.103783.
Kanaani, M., Sedaee, B., Asadian-Pakfar, M., Gilavand, M., Almahmoudi, Z., 2023.
Development of multi-objective co-optimization framework for underground
hydrogen storage and carbon dioxide storage using machine learning algorithms.
J. Clean. Prod. 386, 135785 https://doi.org/10.1016/J.JCLEPRO.2022.135785.
Killough, J.E., 1976. Reservoir simulation with History-dependent saturation functions.
Soc. Petrol. Eng. J. 16, 37–48. https://doi.org/10.2118/5106-PA.
Korrani, A.K., Jerauld, G.R., Sepehrnoori, K., 2016. Mechanistic modeling of low-salinity
waterflooding through coupling a geochemical package with a compositional
reservoir simulator. SPE Reservoir Eval. Eng. 19, 142–162.
Kumar, K.R., Makhmutov, A., Spiers, C.J., Hajibeygi, H., 2021. Geomechanical
simulation of energy storage in salt formations. Sci. Rep. 11, 1–24.
Laban, M., 2020. Hydrogen Storage in Salt Caverns: Chemical Modelling and Analysis of
Large-Scale Hydrogen Storage in Underground Salt Caverns. Delft University of
Technology.
Land, C.S., 1968. Calculation of imbibition relative permeability for two- and three-phase
flow from rock properties. Soc. Petrol. Eng. J. 8, 149–156. https://doi.org/10.2118/
1942-PA.
Lewandowska-Śmierzchalska, J., Tarkowski, R., Uliasz-Misiak, B., 2018. Screening and
ranking framework for underground hydrogen storage site selection in Poland. Int. J.
Hydrogen Energy 43, 4401–4414. https://doi.org/10.1016/j.ijhydene.2018.01.089.
Liu, W., Zuo, H., Wang, J., Xue, Q., Ren, B., Yang, F., 2021. The production and
application of hydrogen in steel industry. Int. J. Hydrogen Energy 46, 10548–10569.
https://doi.org/10.1016/J.IJHYDENE.2020.12.123.
Luboń, K., Tarkowski, R., 2020. Numerical simulation of hydrogen injection and
withdrawal to and from a deep aquifer in NW Poland. Int. J. Hydrogen Energy 45,
2068–2083. https://doi.org/10.1016/J.IJHYDENE.2019.11.055.
Luboń, K., Tarkowski, R., 2021. Influence of capillary threshold pressure and injection
well location on the dynamic CO2 and H2 storage capacity for the deep geological
structure. Int. J. Hydrogen Energy 46, 30048–30060. https://doi.org/10.1016/J.
IJHYDENE.2021.06.119.
Luboń, K., Tarkowski, R., 2023. The influence of the first filling period length and
reservoir level depth on the operation of underground hydrogen storage in a deep
aquifer. Int. J. Hydrogen Energy 48, 1024–1042. https://doi.org/10.1016/J.
IJHYDENE.2022.09.284.
Lysyy, M., Fernø, M., Ersland, G., 2021. Seasonal hydrogen storage in a depleted oil and
gas field. Int. J. Hydrogen Energy 46, 25160–25174. https://doi.org/10.1016/j.
ijhydene.2021.05.030.
Lysyy, M., Føyen, T., Johannesen, E.B., Fernø, M., Ersland, G., 2022a. Hydrogen relative
permeability hysteresis in underground storage. Geophys. Res. Lett. 49,
e2022GL100364 https://doi.org/10.1029/2022GL100364.
M. Saeed and P. Jadhawar
Lysyy, M., Ersland, G., Fernø, M., 2022b. Pore-scale dynamics for underground porous
media hydrogen storage. Adv. Water Resour. 163 https://doi.org/10.1016/j.
advwatres.2022.104167.
Machel, H.G., 2001. Bacterial and thermochemical sulfate reduction in diagenetic
settings — old and new insights. Sediment. Geol. 140, 143–175. https://doi.org/
10.1016/S0037-0738(00)00176-7.
Mahdi, D.S., Al-Khdheeawi, E.A., Yuan, Y., Zhang, Y., Iglauer, S., 2021. Hydrogen
underground storage efficiency in a heterogeneous sandstone reservoir. Advances in
Geo-Energy Research 5, 437–443. https://doi.org/10.46690/AGER.2021.04.08.
McCay, M.H., Shafiee, S., 2020. Hydrogen: an energy carrier. Future Energy: Improved,
Sustainable and Clean Options for Our Planet 475–93. https://doi.org/10.1016/
B978-0-08-102886-5.00022-0.
Muhammed, N.S., Haq, B., Al Shehri, D., Al-Ahmed, A., Rahman, M.M., Zaman, E., 2022.
A review on underground hydrogen storage: Insight into geological sites, influencing
factors and future outlook. Energy Rep. 8, 461–499. https://doi.org/10.1016/J.
EGYR.2021.12.002.
Muhammed, N.S., Haq, M.B., Al Shehri, D.A., Al-Ahmed, A., Rahman, M.M., Zaman, E.,
et al., 2023. Hydrogen storage in depleted gas reservoirs: a comprehensive review.
Fuel 337, 127032. https://doi.org/10.1016/J.FUEL.2022.127032.
Nasrifar, K., 2010. Comparative study of eleven equations of state in predicting the
thermodynamic properties of hydrogen. Int. J. Hydrogen Energy 35, 3802–3811.
https://doi.org/10.1016/J.IJHYDENE.2010.01.032.
Ng, C.S.W., Djema, H., Nait Amar, M., Jahanbani Ghahfarokhi, A., 2022. Modeling
interfacial tension of the hydrogen-brine system using robust machine learning
techniques: Implication for underground hydrogen storage. Int. J. Hydrogen Energy
47, 39595–39605. https://doi.org/10.1016/J.IJHYDENE.2022.09.120.
Pan, B., Liu, K., Ren, B., Zhang, M., Ju, Y., Gu, J., et al., 2023. Impacts of relative
permeability hysteresis, wettability, and injection/withdrawal schemes on
underground hydrogen storage in saline aquifers. Fuel 333. https://doi.org/
10.1016/j.fuel.2022.126516.
Panfilov, M., 2016. Underground and pipeline hydrogen storage. Compendium of
Hydrogen Energy 91–115. https://doi.org/10.1016/B978-1-78242-362-1.00004-3.
Parkhurst, D., Appelo, C.A.J., 2013. Description of Input and Examples for PHREEQC
Version 3: a Computer Program for Speciation, Batch-Reaction, One-Dimensional
Transport, and Inverse Geochemical Calculations. U.S. Geological Survey.
Peng, D.Y., Robinson, D.B., 1976. A new two-constant equation of state. Ind. Eng. Chem.
Fund. 15, 59–64. https://doi.org/10.1021/I160057A011/ASSET/I160057A011.FP.
PNG_V03.
Pfeiffer, W.T., Bauer, S., 2015. Subsurface porous media hydrogen storage - scenario
development and simulation. Energy Proc. 76, 565–572. https://doi.org/10.1016/j.
egypro.2015.07.872.
Pfeiffer, W.T., al Hagrey, S.A., Köhn, D., Rabbel, W., Bauer, S., 2016. Porous media
hydrogen storage at a synthetic, heterogeneous field site: numerical simulation of
storage operation and geophysical monitoring. Environ. Earth Sci. 75, 1–18. https://
doi.org/10.1007/S12665-016-5958-X/FIGURES/9.
Pfeiffer, W.T., Beyer, C., Bauer, S., 2017. Hydrogen storage in a heterogeneous sandstone
formation: Dimensioning and induced hydraulic effects. Petrol. Geosci. 23, 315–326.
https://doi.org/10.1144/PETGEO2016-050/ASSET/3F249536-0B1A-4694-B695-
01E35A0084F6/ASSETS/IMAGES/LARGE/PETGEO2016-050.08.JPG.
Pruess, K., Oldenburg, C.M., Moridis, G.J., 1999. TOUGH2 User’s guide Version 2.
https://doi.org/10.2172/751729.
Rabii, A., Aldin, S., Dahman, Y., Elbeshbishy, E., 2019. A review on Anaerobic Co-
Digestion with a focus on the microbial Populations and the effect of multi-stage
Digester configuration. Energies 12, 1106. https://doi.org/10.3390/EN12061106,
2019;12:1106.
Ruiz Maraggi, L.M., Moscardelli, L.G., 2023. Modeling hydrogen storage capacities,
injection and withdrawal cycles in salt caverns: Introducing the GeoH2 salt storage
and cycling app. Int. J. Hydrogen Energy 48, 26921–26936. https://doi.org/
10.1016/J.IJHYDENE.2023.03.293.
Saeed, M., Jadhawar, P., 2023. Optimizing underground hydrogen storage in aquifers:
the impact of cushion gas type. Int. J. Hydrogen Energy. https://doi.org/10.1016/J.
IJHYDENE.2023.08.352.
Saeed, M., Jadhawar, P., Bagala, S., 2023a. Geochemical effects on storage gases and
reservoir rock during underground hydrogen storage: a depleted North Sea oil
reservoir case study. Hydrogen 4, 323–337. https://doi.org/10.3390/
HYDROGEN4020023, 2023;4:323–37.
Saeed, M., Sayani, J.K.S., Jadhawar, P., 2023b. Evaluating the performance of various
cushion gas types for underground hydrogen storage in an aquifer. In: 1st
International Conference on Green Hydrogen for Global Decarbonisation. Pandit
Deendayal Energy University.
Sainz-Garcia, A.A., 2017. Dynamics of Underground Gas Storage. Insights from
Numerical Models for Carbon Dioxide and Hydrogen.
Sainz-Garcia, A., Abarca, E., Rubi, V., Grandia, F., 2017. Assessment of feasible strategies
for seasonal underground hydrogen storage in a saline aquifer. Int. J. Hydrogen
Energy 42, 16657–16666. https://doi.org/10.1016/J.IJHYDENE.2017.05.076.
Sambo, C., Dudun, A., Samuel, S.A., Esenenjor, P., Muhammed, N.S., Haq, B., 2022.
A review on worldwide underground hydrogen storage operating and potential
fields. Int. J. Hydrogen Energy 47, 22840–22880. https://doi.org/10.1016/J.
IJHYDENE.2022.05.126.
Sandler, S.I., 2006. Chemical, Biochemical, and Engineering Thermodynamics 2, 960.
26
Gas Science and Engineering 121 (2024) 205196
Sass, R., 2013. Sources of Thermophysical Properties for Efficient Use in Product Design.
In: Product Design and Engineering: Formulation of Gels and Pastes, vols. 153–67.
https://doi.org/10.1002/9783527654741.CH6.
Schlumberger Ltd, 2014. Eclipse Reference Manual.
Singla, M.K., Nijhawan, P., Oberoi, A.S., 2021. Hydrogen fuel and fuel cell technology for
cleaner future: a review. Environ. Sci. Pollut. Control Ser. 28, 15607–15626. https://
doi.org/10.1007/S11356-020-12231-8, 2021 28:13.
Tarkowski, R., 2019. Underground hydrogen storage: characteristics and prospects.
Renew. Sustain. Energy Rev. 105, 86–94. https://doi.org/10.1016/J.
RSER.2019.01.051.
Thaysen, E.M., McMahon, S., Strobel, G.J., Butler, I.B., Ngwenya, B.T., Heinemann, N.,
et al., 2021. Estimating microbial growth and hydrogen consumption in hydrogen
storage in porous media. Renew. Sustain. Energy Rev. 151 https://doi.org/10.1016/
j.rser.2021.111481.
Tremosa, J., Jakobsen, R., Le Gallo, Y., 2023. Assessing and modeling hydrogen
reactivity in underground hydrogen storage: a review and models simulating the
Lobodice town gas storage. Front. Energy Res. 11, 1145978 https://doi.org/
10.3389/FENRG.2023.1145978/BIBTEX.
van Rooijen, W.A., Habibi, P., Xu, K., Dey, P., Vlugt, T.J.H., Hajibeygi, H., et al., 2022.
Interfacial tensions, solubilities, and transport properties of the H2/H2O/NaCl
system: a molecular simulation study. J. Chem. Eng. Data. https://doi.org/10.1021/
ACS.JCED.2C00707/ASSET/IMAGES/LARGE/JE2C00707_0005.JPEG.
Vo Thanh, H., Rahimi, M., Dai, Z., Zhang, H., Zhang, T., 2023. Predicting the wettability
rocks/minerals-brine-hydrogen system for hydrogen storage: Re-evaluation
approach by multi-machine learning scheme. Fuel 345, 128183. https://doi.org/
10.1016/J.FUEL.2023.128183.
Wang, G., Pickup, G., Sorbie, K., Mackay, E., 2022a. Scaling analysis of hydrogen flow
with carbon dioxide cushion gas in subsurface heterogeneous porous media. Int. J.
Hydrogen Energy 47, 1752–1764. https://doi.org/10.1016/J.
IJHYDENE.2021.10.224.
Wang, G., Pickup, G., Sorbie, K., Mackay, E., 2022b. Numerical modelling of H2 storage
with cushion gas of CO2 in subsurface porous media: Filter effects of CO2 solubility.
Int. J. Hydrogen Energy 47, 28956–28968. https://doi.org/10.1016/J.
IJHYDENE.2022.06.201.
Wang, H., Xin, Y., Kou, Z., Qu, Y., Wang, L., Ning, Y., et al., 2023a. Numerical study of
the efficiency of underground hydrogen storage in deep saline aquifers, rock springs
uplift, Wyoming. J. Clean. Prod. 421, 138484 https://doi.org/10.1016/J.
JCLEPRO.2023.138484.
Wang, S., Chen, F., Wu, Y.-S., Nasrabadi, H., 2023b. The potential of hydrogen storage in
depleted unconventional gas reservoirs: a multiscale modeling study. Int. J.
Hydrogen Energy. https://doi.org/10.1016/J.IJHYDENE.2023.01.092.
Wang, J., Yang, Y., Cai, S., Yao, J., Xie, Q., 2023c. Pore-scale modelling on hydrogen
transport in porous media: implications for hydrogen storage in saline aquifers. Int.
J. Hydrogen Energy 48, 13922–13933. https://doi.org/10.1016/J.
IJHYDENE.2022.11.299.
Wiebe, R., Gaddy, V.L., 1934. The solubility of hydrogen in water at 0, 50, 75 and 100◦
from 25 to 1000 atmospheres. J. Am. Chem. Soc. 56, 76–79. https://doi.org/
10.1021/JA01316A022.
Yekta, A.E., Pichavant, M., Audigane, P., 2018. Evaluation of geochemical reactivity of
hydrogen in sandstone: application to geological storage. Appl. Geochem. 95,
182–194. https://doi.org/10.1016/J.APGEOCHEM.2018.05.021.
Yu, M., Wang, K., Vredenburg, H., 2021. Insights into low-carbon hydrogen production
methods: green, blue and aqua hydrogen. Int. J. Hydrogen Energy 46, 21261–21273.
https://doi.org/10.1016/J.IJHYDENE.2021.04.016.
Zeng, L., Keshavarz, A., Xie, Q., Iglauer, S., 2022a. Hydrogen storage in Majiagou
carbonate reservoir in China: geochemical modelling on carbonate dissolution and
hydrogen loss. Int. J. Hydrogen Energy 47, 24861–24870. https://doi.org/10.1016/
j.ijhydene.2022.05.247.
Zeng, L., Hosseini, M., Keshavarz, A., Iglauer, S., Lu, Y., Xie, Q., 2022b. Hydrogen
wettability in carbonate reservoirs: Implication for underground hydrogen storage
from geochemical perspective. Int. J. Hydrogen Energy 47, 25357–25366. https://
doi.org/10.1016/j.ijhydene.2022.05.289.
Zeng, L., Vialle, S., Ennis-King, J., Esteban, L., Sarmadivaleh, M., Sarout, J., et al., 2023.
Role of geochemical reactions on caprock integrity during underground hydrogen
storage. J. Energy Storage 65, 107414. https://doi.org/10.1016/J.
EST.2023.107414.
Zhan, S., Zeng, L., Al-Yaseri, A., Sarmadivaleh, M., Xie, Q., 2023. Geochemical modelling
on the role of redox reactions during hydrogen underground storage in porous
media. Int. J. Hydrogen Energy. https://doi.org/10.1016/j.ijhydene.2023.06.153.
Zhang, J., Clennell, M.B., Sagotra, A., Pascual, R., 2023. Molecular dynamics simulation
and machine learning for predicting hydrogen solubility in water: effects of
temperature, pressure, finite system size and choice of molecular force fields. Chem.
Phys. 564, 111725 https://doi.org/10.1016/J.CHEMPHYS.2022.111725.
Zhao, Q., Wang, H., Chen, C., 2024. Underground hydrogen storage: a recovery
prediction using pore network modeling and machine learning. Fuel 357, 130051.
https://doi.org/10.1016/J.FUEL.2023.130051.
Zhu, Y., Wang, H., Vano, K., 2022. Applying the wavelet neural network to estimate
hydrogen dissolution in underground sodium chloride solutions. Int. J. Hydrogen
Energy 47, 22720–22730. https://doi.org/10.1016/J.IJHYDENE.2022.05.130.
Zivar, D., Kumar, S., Foroozesh, J., 2021. Underground hydrogen storage: a
comprehensive review. Int. J. Hydrogen Energy 46, 23436–23462. https://doi.org/
10.1016/J.IJHYDENE.2020.08.138.