Impacts of reservoir permeability anisotropy on hydrogen migration and recovery efficiency in geological storage

The problem

Surplus wind and solar electricity can be converted to hydrogen, stored underground, and turned back into electricity when demand exceeds generation (Ge et al., 2024; Berna-Escriche et al., 2022). At standard surface conditions, hydrogen holds about a third of the energy of methane per unit volume (Poling et al., 2020). Depleted gas fields have large pore space and existing infrastructure, and their seal has already held gas (Harati et al., 2023).

A key measure of a storage site is its hydrogen recovery efficiency, the share of the injected hydrogen that is withdrawn again. Permeability, the ease with which fluid flows through rock, often differs with direction. Ratios of horizontal to vertical permeability (kh/kv) are typically 1 to 100 (e.g. Goupil et al., 2022). Earlier studies of hydrogen storage mainly examined kh/kv (Ershadnia et al., 2023; Camargo et al., 2024). In this study, we evaluated both kh/kv and the ratio of the largest to the smallest horizontal permeability (kh,max/kh,min).

What we did

We built a box-shaped model of a depleted gas reservoir, 1000 m by 1000 m (one square kilometre) and 50 m thick. In the baseline, permeability was 100 mD (millidarcies) in every direction. One vertical well at the centre was open over the full thickness. We first injected natural gas for 360 days as cushion gas. Ten cycles followed, each with 90 days of hydrogen injection and 90 days of withdrawal at 1 million Sm³ (standard cubic metres) per day. We used the simulator OpenGoSim and neglected gas dissolution in brine, chemical and biological reactions, and capillary pressure.

We lowered vertical permeability to 20, 10 and 1 mD for kh/kv of 5, 10 and 100. For kh,max/kh,min of 5 and 10, we kept the geometric mean of the two horizontal permeabilities at 100 mD. At a ratio of 10, permeability was 316 mD one way and 32 mD across it. We combined these cases with thicknesses of 40 and 60 m, average permeabilities of 50 and 200 mD, and six well layouts. We then repeated the tests in a model of the Ahuroa field in Taranaki, a former gas field that has stored natural gas since 2008. The hydrogen scenario there is hypothetical.

What we found

In our box model, lower vertical permeability gave higher recovery. As kh/kv increased from 1 to 100, recovery after ten cycles went from 74% to 82%. This is because hydrogen is lighter than natural gas. It rises and spreads along the top of the reservoir, a process called gravity override. Low vertical permeability limited this and retained more hydrogen in the lower layers. With similar gas rates from every layer, the well withdrew gas richer in hydrogen.

The effect of horizontal anisotropy was smaller and increased with average permeability. A kh,max/kh,min of 10 raised recovery by about 1 percentage point at 100 mD, and from 64% to 68% at 200 mD. At 50 mD it lowered recovery from 81% to 80%. Higher permeability on its own reduced recovery, but a kh,max/kh,min of 10 may offset part of that loss. A higher kh/kv weakened the horizontal effect, to at most 1 percentage point at a kh/kv of 100.

Anisotropy also reduced the effect of reservoir thickness on recovery. With no anisotropy, a 60 m reservoir recovered 5 percentage points less than a 40 m one. At the end, the lowest 10 m or so of the 60 m reservoir held less than 10% hydrogen, against the lowest 2 m of the 40 m reservoir. This larger low-hydrogen zone at the base contributes to the lower recovery. The difference fell to 3 points at a kh,max/kh,min of 10, and to 1 point at a kh/kv of 100. We attribute this to anisotropy restricting gravity override.

The explorer below gives box-model recovery for every combination of average permeability, kh/kv and kh,max/kh,min.

Well orientation also affected recovery. At a kh,max/kh,min of 10, with a vertical injection well and a horizontal withdrawal well, recovery was 78% with the horizontal well along the most permeable direction and 70% along the least permeable one. For a single horizontal well used for both injection and withdrawal, the values were 77% and 74%.

High vertical anisotropy favoured wells open to the reservoir only near its top. At a kh/kv of 100, these gave 84 to 85%, against 78% for the vertical injection and horizontal withdrawal layout. This is because hydrogen injected into the lower layers met resistance to upward flow, with a vertical permeability of only 1 mD. We do not recommend that layout where kh/kv is high.

In the Ahuroa model, recovery increased from 76.8% to 82.8% as kh/kv went from 1 to 100, and to 77.3% as kh,max/kh,min went from 1 to 10. Thin low-permeability interlayers also make a whole reservoir anisotropic. When the permeability of one interlayer rock type fell from 10 to 0.1 mD, the reservoir-scale kh/kv rose from 2.6 to 167, and recovery from 75.2% to 79.3%.

Limitations

We assumed the most permeable direction was the same in every layer, whereas permeability usually varies between strata (Zhao et al., 2024; Tarkowski et al., 2024). We did not include spatial heterogeneity, the variation of permeability from place to place. Every result here is from simulations.

For anisotropy, laboratory measurements on core samples may not be enough because of their small scale (Raghavan, 2006). We argue that site assessments for hydrogen storage need a reservoir-scale estimate of anisotropy, in addition to core measurements of single layers.

Glossary

Permeability: the ease with which a fluid flows through rock under a pressure difference. Millidarcy (mD): a unit of permeability. At Ahuroa, the median permeability of the main rock types ranges from 0.01 mD in mudstone to 75 mD in the main sandstone. Permeability anisotropy: a difference in permeability between directions. kh/kv: the ratio of horizontal to vertical permeability. kh,max/kh,min: the ratio of the largest to the smallest horizontal permeability. Geometric mean: the square root of the product of two values, used here as the average horizontal permeability. Hydrogen recovery efficiency: the share of injected hydrogen that is withdrawn again. Cushion gas: gas left in the reservoir to maintain reservoir pressure during withdrawal, here natural gas. Gravity override: the rise of light hydrogen to the top of the reservoir, where it spreads sideways. Sm³: standard cubic metre, a gas volume measured at surface conditions. Core sample: a cylinder of rock cut from a borehole for laboratory testing.