The storage of hydrogen gas in underground lined rock caverns (LRCs) enables the implementation of the first fossil-free steelmaking process to meet the large demand for crude steel. Predicting the response of rock mass is important to ensure that gas leakage due to rupture of the steel lining does not occur. Analytical and numerical models can be used to estimate the rock mass response to high internal pressure; however, the fitness of these models under different in situ stress conditions and cavern shapes has not been studied. In this paper, the suitability of analytical and numerical models to estimate the maximum cavern wall tangential strain under high internal pressure is studied. The analytical model is derived in detail and finite element (FE) models considering both two-dimensional (2D) and three-dimensional (3D) geometries are presented. These models are verified with field measurements from the LRC in Skallen, southwestern Sweden. The analytical model is inexpensive to implement and gives good results for isotropic in situ stress conditions and large cavern heights. For the case of anisotropic horizontal in situ stresses, as the conditions in Skallen, the 3D FE model is the best approach.

The storage of hydrogen gas in lined rock caverns (LRCs) may enable the implementation of the first large-scale fossil-free steelmaking process in Sweden, but filling such storage causes joints in the rock mass to open, concentrating strains in the lining. The structural interaction between the LRC components must be able to reduce the strain concentration in the sealing steel lining; however, this interaction is complex and difficult to predict with analytical methods. In this paper, the strain concentration in LRCs from the opening of rock joints is studied using finite element (FE) analyses, where the large- and small-scale behaviors of the LRC are coupled. The model also includes concrete crack initiation and development with increasing gas pressure and rock joint width. The interaction between the jointed rock mass and the reinforced concrete, the sliding layer, and the steel lining is demonstrated. The results show that the rock mass quality and the spacing of the rock joints have the greatest influence on the strain distributions in the steel lining. The largest effect of rock joints on the maximum strains in the steel lining was observed for geological conditions of “good” quality rock masses.

Efforts to substitute the use of fossil fuels in industry by hydrogen gas requires the storage of large volumes of gas with a reliable pressure vessel design. The Hydrogen Breakthrough Ironmaking Technology (HYBRIT) initiative aims to make the whole steel making process in Sweden fossil-free with the storage of industrial scale quantities of hydrogen in underground Lined Rock Cavers (LRCs). The LRC concept is a relatively new design methodology that can be further developed with respect to safety and economic efficiency and reliability-based design methods provide one option to comply with codes and regulations. High reliability is required for the storage of hydrogen gas and the computational time becomes unpractical for the evaluation of a complex system such as the LRC. In this paper, the efficiency of Subset Simulation (SuS) regarding accuracy, precision and required number of samples is studied for the calculation of probability of failure against fatigue of the steel lining. It can be observed that by increasing the number of samples per level and increasing the conditional probability of failure the precision increases as well as the total number of samples. The accuracy of the SuS is checked with respect to Monte Carlo simulation (MCS) showing good agreement and with greater precision for fewer number of samples. A case study is performed for the geologic conditions of Sweden showing that the considered failure mode is unlikely for high stresses and good rock mass quality.

Efforts to decrease CO2 emissions in the Swedish steelmaking process involve the use of industrial quantities of hydrogen gas supplied from large-scale Lined Rock Cavern (LRC) storages in order to eliminate the use of fossil fuels. This storage must be placed at sufficient depth so that the overburden is able to resist the uplifting gas pressure from inside the cavern. Both the high reliability requirement and addressing the uncertainties related to the rock mass properties make it difficult to design for such structures. In this work, a reliability-based design methodology for the LRC depth specification using the Response surface (RS) method is presented. Geologic conditions of Sweden, i.e. hard rock, are considered and the analytical solution for the resistance to uplift includes the tensile strength of the failure surface in addition to the overburden weight pressure. The highest uncertainties are assumed to be related to the rock mass parameters and both the cavern radius and the maximum operational pressure are chosen to be the same as for the LRC in Skallen, in southwestern Sweden. Four random variables with varying correlation are used to estimate the acceptable cavern depth and the results are reasonable compared to previous experience. The efficiency of the RS method for the considered problem is observed both for required number of samples and accuracy, showing suitability to be used with more complex, difficult to evaluate, problems such as Finite Element models.

The transition to a fossil-free energy matrix may require large quantities of hydrogen gas, which could be stored efficiently in an underground lined rock cavern (LRC). Since the consequences of failure can be catastrophic, the LRC design needs to have a small probability of failure. However, the current design practice for LRCs is deterministic, which limits the possibility to stringently address geotechnical uncertainties in the design. In this paper, a reliability-based design tool is presented for LRCs. The adaptive directional importance sampling (ADIS) method, which requires a relatively small number of samples, is used with a 3D finite element (FE) model to evaluate small probabilities of failure. An illustrative example based on the LRC in Skallen, southwestern Sweden, demonstrates the implementation and applicability of the developed design tool. The considered uncertainties are related to the geological conditions and the steel lining. The results show that the reliability of this LRC design meets the expected safety requirements. Considering different geological conditions with correlations, at least “good” quality rock mass is needed for the LRC design. An additional sensitivity analysis is performed to study the potential influence of corrosion and hydrogen embrittlement on the reduction of the LRC design reliability.