Project Topics
Get Complete Project Material File(s) Now! »
Hydrodynamic behavior of underground hydrogen storage
Hydrogen was discovered by Henry Cavendish in 1766, is the highly abundant and the main element (about 75%) in the universe and has the second lowest melting (13.99K) and boiling points (20.271K) at atmospheric pressure (Figure 2-4) with only Helium being below. This is one of the reasons why hydrogen is not used as a primary fuel. Indeed, hydrogen is more difficult to store under standard conditions compared to other gases which can be liquefied at standard temperature. The boiling point of hydrogen can only be increased to 20K peaking at a pressure of 13 bar. However, at standard temperature and pressure conditions, hydrogen is a colorless, odorless, tasteless, nontoxic, noncorrosive, nonmetallic which is in principle physiologically not dangerous. One of its most important characteristics is its low density (0.084 Kg/m3 at surface conditions – Mallard et al., 1998 ) making it less dense than air, which makes it necessary for any practical applications to either compress the hydrogen or liquefy it. However, hydrogen gas has one of the widest flammability ranges of concentrations between 4 percent and 75 percent by volume (Lanz et al., 2001). Due to its high molecular velocity, hydrogen also has the highest diffusivity of any gas. For instance, hydrogen molecules disperse in air four times faster than molecules of natural gas. Thus, if hydrogen does leak, it will disperse rapidly. This has very important implications for managing hydrogen safety. Hydrogen is only dangerous if it is released in rooms where it can accumulate to explosive mixtures. In open areas, outdoors or in a large area, even big leaks do not pose a threat, because of the high buoyancy and diffusivity of hydrogen allows to rise rapidly, moving away from potential ignition sources, personnel, or other equipment. Although hydrogen is highly reactive and can be ignited very easily, the risk of spontaneous ignition is low because of its auto ignition temperature is 585°C (NIST 2012).
Hydrogen has a specific energy density of 33.3 kWh/kg or ~124 MJ/kg (Carden and Paterson 1979) which offers a unique potential to store large amounts of energy (Zittel et al., 1996). Compared to methane (representing natural gas), hydrogen has roughly 1/3 smaller energy content per m3. Each m3 of hydrogen produces 12.7 MJ of energy by combustion (Züttel 2004) instead of 40 MJ for methane. Thus, energy potential of hydrogen is not enough to be considered as a primary source of energy such as petroleum (Marbán et al., 2007).
Hydrogen Water fluid flow equations
The classical mass conservation in terms of mass fraction 𝑋𝛼,𝑘 of species 𝑘 in fluid phase 𝛼 is (Bear 1972): 𝜕(∅Σ𝑆𝛼𝜌𝛼𝑋𝛼,𝑘𝛼)𝜕𝑡+ ∇.Σ(𝜌𝛼𝑋𝛼,𝑘𝑣𝛼+𝐽𝛼,𝑘)𝛼=𝑅𝛼,𝑘+𝑄𝛼,𝑘 (2-1).
where ∅ is the porosity, 𝑅𝛼,𝑘 is the reaction term, 𝑄𝛼,𝑘 is the source term, 𝜌𝛼 is the molar density, 𝑆𝛼 is the phase saturation, 𝑣𝛼 is the Darcy velocity of fluid phase, 𝐽𝛼,𝑘 is the total diffusive and dispersive flux. Moreover, below equations are considered for the phase saturations and concentrations sum to 1:
Σ𝑆𝛼𝛼=1 (2-2).
Σ𝑋𝛼,𝑘𝑘=1 (2-3).
The momentum balance for each phase at macroscale is formulated with the multi-phase Darcy law: 𝑣𝛼=−𝐾𝑘𝑟𝛼𝜇𝛼(∇𝑃𝛼−𝜌𝛼𝑔) (2-4).
where K is the absolute permeability, 𝑘𝑟𝛼 is the relative permeability, 𝑃𝛼 is the phase pressure, 𝜇𝛼 is the phase viscosity and 𝑔 is the gravity. Phase pressures are correlated by capillary pressure, for instance for water and the gas: 𝑃𝑐(𝑆𝑤)=𝑃𝑔−𝑃𝑤 (2-5).
Hydraulic properties of porous media (capillary pressure and relative permeabilities of water and hydrogen) are dependent to water saturation, which they could be measured experimentally (see section 5).
Thermodynamics and geochemical fluid-rock modeling
In general, geochemical reactive modeling is based on the primary species formulation, therefore, the equilibrium chemical reactions between the primary and secondary species take the following form (see Steefel et al., 2015 and references in this paper review): 𝐴𝑗⇌Σ𝛽𝑖𝑗𝐴𝑖𝑁𝑐𝑖=1 (2-11). where 𝐴𝑖 and 𝐴𝑗 are the chemical formulas of the primary and secondary species, respectively, 𝛽𝑖𝑗is the number of moles of primary species 𝑖 in one mole of secondary species 𝑗 and 𝑁𝑐 is the number of independent chemical components in the system. The primary and secondary species are linked by the equilibrium reaction via the law of mass action for each reaction: 𝐴𝑗=𝐾𝑗𝛾𝑗−1Π(𝛾𝑖𝐴𝑖)𝜐𝑖𝑗𝑁𝑐𝑖=1 𝑖=1,…,𝑁𝑥 (2-12).
where 𝛾𝑖−1 and 𝛾𝑗 are the activity coefficients for the primary and secondary species, respectively, 𝐾𝑗 is the reaction equilibrium constant and 𝑁𝑥 is the total number of species. Therefore, the total concentrations is: 𝜓𝑗=𝐴𝑗+Σ𝛽𝑖𝑗𝐴𝑗𝑁𝑥𝑗=1 (2-13).
Density and viscosity of the hydrogen and water are correlated with respect to the pressure, temperature and phase composition. However, the compositional flow system (Equation (2-1)) has to be closed by the phase equilibria conditions. The equilibrium criterion (equilibrium between hydrogen and water), dictates a minimum of the Gibbs free energy at constant temperature T, pressure P and composition: 𝐺=Η−ΤS.
Hydrogen issues on porous storage
The most important difference between storage of natural gas and storage of hydrogen (regardless to hydrogen biotic and abiotic reactions) is related to the properties of hydrogen. Hydrogen has a lower density and lower viscosity than other gases that influence the subsurface behavior and the reservoir performance. Moreover, hydrogen has the lower volumetric heating value than natural gas that impact the energy storage capacity, injection and withdrawal rates and also within the reservoir process.
Hydrogen geochemical interactions
As mentioned before, porous rock storages like deep saline aquifer or depleted oil and gas reservoir are the most promising hydrogen geological storage options on the regional to global scale based on their estimated storage capacities and their widespread distribution. However, the injection of hydrogen into these porous rock storages disturb the initial equilibrium and could be caused chemical interactions between injected hydrogen, saline formation fluid and reservoir rock. These interactions include dissolution of minerals and precipitation of others (Ganzer et al., 2013; Truche et al., 2013) and not only change the chemical, but rather the physical properties of the reservoir system such as permeability, porosity or injectivity. The chemical interactions may lead to mobilization of initial components while the changes in physical properties influence operation, storage capacity, as well as long-term safety and stability of the reservoir. Hence, precise knowledge of the hydrogen-induced interactions between injected hydrogen and reservoir rocks and the resulting changes in chemical and physical properties of the reservoir system is therefore a prerequisite for any secure operation of a storage site.
Various supplementary methods can be undertaken to evaluate the behavior of the gas stored (like CO2 or Hydrogen storage) in sedimentary formation and its interactions with water and the minerals of the host formation in a storage site.
Field study is the fundamental method to characterize of natural gas storage sites, which makes it possible to consider geological timescales at the reservoir scales. This method maybe not possible for hydrogen storage due to the absence of natural hydrogen storage and the investigation of hydrogen storage on sedimentary formation returns to recent years. Other methods are laboratory experiments and numerical simulations. Laboratory experiments provide direct observations of gas-fluid-rock interactions, however, at the experiment duration and spatial scales. This method is applicable for hydrogen gaseous that interaction of hydrogen with components of rock mineral compositions at the experiment time scales can be investigated. The last method is numerical simulations to interpret the geochemical reactivity behavior of a hydrogen storage site. This method include overlarge time and spatial scales that for gas storage modeling requires a good description of the process taking place and a precise characterization of the parameters involved in the calculations.
Laboratory experiments support numerical method with rate data for a large number of minerals at various temperatures, solute concentrations and at various distances from equilibrium. Thus, numerical modeling requires information from laboratory experiments and without any laboratory experiments and valid data, the results from numerical simulation not reliable.
However, laboratory experiments at simulated reservoir pressure and temperature conditions are an elegant way to study the hydrogen-water-rock interactions. Despite the importance of this study for underground hydrogen storage, unfortunately, experimental analysis at the reservoir conditions are scarce. Therefore, further experimental studies (either focus on individual minerals or on whole rock samples) that cover additional physico-chemical conditions (e.g., pressure, temperature, lithology, brine composition), refer to the problem of potential slow reaction kinetics by needed prolonged run.
Hydrogen through the water dissolution reaction has a strong reduction power in a chemical system (Lassin et al., 2011). Molecule of hydrogen has a polar nature and the strong H-H binding energy (436 kJ/mol) requires the overstepping of a high energetic barrier (Truche et al., 2013). Thus, most of the possible redox reactions induced by hydrogen remain insignificant at low temperature, even on a geological time scale, provided no bacteria is present. In general, two types of reactions induced by hydrogen in the underground storage could be considered: abiotic and biotic reactions.
Abiotic reactions
The mechanisms and kinetics of redox reactions induced by hydrogen on confining rocks are yet poorly documented. However, recently abiotic hydrogen reactivity with rock minerals, which are restricted to redox reactions at reservoir temperature, has been taken into consideration. Truche et al,. (2013) studied experimentally abiotic redox reaction induced by hydrogen at low temperature. However, he has illustrated Pyrite mineral reduction into Pyrrhotite and releasing sulfide anions in the solution (Equation (2-40)), under hydrogen partial pressure above 30 bar and temperature as high as 150°C, which is higher for the underground hydrogen storage. In addition, he has demonstrated that the alkalinity of the geological storage impact the abiotic redox reaction and the pH of the media is a critical parameter to control the extension of the reaction at low temperature. Hence, pH changes can also lead to mineralogical transformations. 𝐹𝑒𝑆2+(1−𝑥)𝐻2=𝐹𝑒𝑆1+𝑥+(1−𝑥)𝐻2𝑆 ,0≤𝑥<0.125 (2-40).
However, injecting hydrogen in the porous formation storage can promote abiotic reaction of hydrogen with minerals of host reservoir and caprocks that could be caused dissolution of carbonate, sulfate, feldspars and clay minerals and also precipitation of illite, iron sulfide and pyrrhotite (Reitenbach et al., 2015) which should be approved experimentally.
Biotic reactions
Besides the abiotic reactions, at underground hydrogen storage the microbial activities that could cause hydrogen consumption is the most likely phenomenon, which can affect the geochemical environment of a gas storage and lead to a loss of hydrogen (Reitenbach et al., 2015). In fact, bacteria consume the energy produced from redox reaction that initiated from hydrogen and the other components in reservoir, while they do not consume hydrogen directly (Panfilov 2016). However, the evidences from town gas storages in Lobodice (Czech Republic) and also in Beynes (France) reveal the biotic reactions and bacteria activity in hydrogen storage.
Rock core samples and analytical methods
The purpose of this study was the testing of sandstone lithologies for underground hydrogen storage. Therefore, lower Triassic sandstones from the Buntsandstein (Lower Triassic) formation east of the Paris Basin were chosen. The Paris Basin is the largest on-shore French sedimentary basin (Bader et al., 2014). It has been identified as a major site for geothermal storage (Aquilina et al., 1997; Blaise et al., 2016). The Buntsandstein was deposited under calm tectonic conditions and represents the main subsidence period of the eastern layer of the Paris Basin. Therefore, deposits exhibit uniform lithologies and thicknesses over large distances. The Buntsandstein represents the lower group of the Trias (Figure 3-1). It is subdivided by lithological criteria into three levels: (a) Voltzia sandstone, (b) Couches intermediate sandstone, and (c) the Vosges sandstone.
Intrinsic permeability measurements
Intrinsic or absolute permeability specifies the ability of a fluid to penetrate through a rock. Intrinsic permeability is defined by Darcy’s law (Equation (4-1)) in single-phase flow in porous media which is written after Equation (2-4) with one fluid (water or gas), which means 𝑘𝑟,𝑖=1 (gravity is not considered): −Δ𝑃𝐿=𝜇𝐾𝑄𝐴 (4-1).
where 𝑄 is the flow rate of the fluid across a cylindrical core, 𝐴 is the cross-section of the core, 𝐾 is the intrinsic permeability of the core, 𝜇 is the viscosity of the fluid, 𝐿 is the length of the core and Δ𝑃 is the pressure difference across the rock core.
In this study, intrinsic permeability measurements were performed by using single-phase flow of water, hydrogen and of argon gas. Fluid (water or gas) was injected in through the core at several flow rates and after stability, the pressure drop at both ends of the core were recorded. The slope of differential pressure vs. flow rate gives the intrinsic or absolute permeability, taking into account the viscosity of fluid and the geometry of the core.
To measure the intrinsic permeability, just two metering pumps were used, one to inject water or gas into the core sample and the other to collect water or gas (Figure 4-7). In this measurement, the separator is not needed. The first metering pump was set in constant flow rate mode (in this case, the pressure of injection, which is unknown, becomes constant after a while). The second metering pump was set in constant pressure mode (defining the backpressure, ideally in the range of typical reservoir fluid pressures).
Table of contents :
1. General introduction
1.1 Energy storage technologies
1.2 Underground geological storage
1.3 Objectives of the thesis
2. Underground hydrogen storage
2.1 Types of underground hydrogen storage
2.2 Hydrodynamic behavior of underground hydrogen storage
2.2.1 Hydrogen Water fluid flow equations
2.2.2 Thermodynamics and geochemical fluid-rock modeling
2.3 Hydrogen issues on porous storage
2.3.1 Solubility
2.3.2 Viscous instability
2.3.3 Gravity Overriding
2.3.4 Diffusion
2.3.5 Oxidation-Reduction (RedOx) potential
2.4 Hydrogen geochemical interactions
2.4.1 Abiotic reactions
2.4.2 Biotic reactions
3. Experimental study of geochemical reactivity of hydrogen in sandstone
3.1 Rock core samples and analytical methods
3.2 Experimental methods and procedure
3.3 Experimental results
3.4 Discussion
4. Experimental determination of relative permeability and capillary pressure in the hydrogen-water system
4.1 Experimental setup and apparatus
4.2 Experimental conditions and procedures
4.2.1 Intrinsic permeability measurements
4.2.2 Protocol followed for saturation measurements
4.2.3 Capillary pressure measurement
4.2.4 Relative permeability measurement
4.3 Experimental results
4.3.1 Absolute permeability
4.3.2 Capillary pressure measurement
4.3.3 Steady state relative permeability measurements
4.4 Discussion
4.4.1 Discussion of the results
4.4.2 Validation of the core flooding experimental set-up
4.4.3 Capillary end effect
5. Numerical simulation
5.1 Hydrogen-water-rock interaction numerical simulation
5.1.1 Results of geochemical simulations
5.1.2 Temporal evolution of sandstone reservoir in presence of hydrogen
5.2 Numerical simulation of a hydrogen geological storage in the Trias Formation, France
5.2.1 Introduction to underground energy storage
5.2.2 Energy supply in France
5.2.3 Renewables energy in France
5.2.4 Geological potential in France for energy storage: sitology and needs
5.2.5 Modeling of underground storage of hydrogen to compensate a week-long shortage of energy production in Ile de France
5.2.6 Geological model
5.2.7 Geochemical data
5.2.8 Hydrogen storage reactive transport modeling
5.2.9 Results
5.2.10 Discussion
6. General conclusion and applications
7. Perspective
8. Results of papers
8.1 Paper I
8.2 Paper II
References .
