Long-term stability forecasting for energy storage salt caverns using deep learning-based model

Kai Zhao, Sihao Yu and Louis Ngai Yuen Wong

Energy, 2025, vol. 319, issue C

Abstract: Underground salt caverns are widely used for energy storage due to their favorable rheology, low permeability, and self-healing properties after sustaining damage. Ensuring the long-term stability of these caverns is crucial for the safe operation of such projects. However, current evaluation methods are laborious, time-consuming and involve a series of laboratory tests, the establishment of constitutive models, and numerical simulations that model actual in situ formations. This paper proposes a highly efficient deep learning-based method for predicting the long-term stability of energy storage salt caverns. Twelve critical parameters, including cavern geometry, operational parameters, stratigraphic distribution, and mechanical properties of rock cores, were set as input variables. Two parameters, the maximum displacement on the cavern wall and the reduced cavern volume, were set as output variables. An artificial neural network model for salt cavern long-term stability was trained on this dataset. The coefficient of determination (R2) and root mean square error (RMSE) were selected as evaluation metrics. The R2 values for the maximum displacement and reduced cavern volume in both the training and test sets exceeded 0.97. This high level of accuracy meets engineering design requirements, suggesting that the deep learning model can be a potentially suitable tool for predicting the long-term stability of energy storage salt caverns under various conditions. This pioneering work represents the first application of a deep learning model to predict the long-term stability of underground salt caverns. Our novel approach enhances calculation efficiency and provides critical insights for site selection and operational pressure design during the initial phases of large-scale salt cavern gas storage projects.

Keywords: Underground salt cavern; Long-term stability; Deep learning; Artificial neural network (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544225004967
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:319:y:2025:i:c:s0360544225004967

DOI: 10.1016/j.energy.2025.134854

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().