AI-driven revolution in subsurface gas storage: Addressing operational and environmental challenges

Artificial intelligence (AI), particularly machine learning (ML) models such as artificial neural networks (ANNs), support vector machines (SVMs), and deep learning architectures, is revolutionizing subsurface gas storage (SGS) by enabling predictive modelling and operational optimization. This study applies a hybrid AI approach integrating deep learning with geostatistical reservoir simulations to assess operational performance and environmental risks. A case study involving a depleted gas reservoir in Western Europe demonstrates the methodology's effectiveness. Results show that AI-assisted models improve storage capacity prediction accuracy by 18 %, reduce leakage detection time by 35 %, and enhance injection optimization by 22 % compared to conventional methods. Integrating real-time pressure, temperature, and geophysical data with AI models improves risk assessment reliability. These findings underscore the value of AI in supporting safer, more efficient hydrogen and natural gas storage systems. The proposed framework offers a scalable solution adaptable to various geologic settings and storage technologies.