AI-driven microbial sequencing analysis for hydrogen storage in the Maari reservoir

Abstract

Hydrogen storage in subsurface reservoirs has been of great interest in order to overcome seasonal demand and supply discrepancies. Hydrogen is a good energy carrier, whose combustion process solely leads to energy and water as output and can be utilized as an energy storage carrier. Hydrogen’s high energy density can be very beneficial as a reactant for several chemical processes. Given the growing utilization of hydrogen for supporting the energy transition, hydrogen can be stored in various ways, such as metal tanks and in underground subsurface environments. Furthermore, hydrogen may be stored in the form of a mixture of natural gas, materials and in deep geological structures. Underground natural gas storage has been conducted for a long time in porous rock formations and salt caverns and given the massive amounts of hydrogen that have to be stored, subsurface hydrogen storage has been of significant interest. Given the limited experience for underground hydrogen storage and most of the hydrogen storage experience has been in salt caverns. We have developed a data-driven framework for the analysis of microbial effects on subsurface hydrogen storage. The framework integrates 16S rRNA sequencing data from subsurface hydrogen storage sites to analyze their composition and deduce potential effects on the hydrogen productivity from the reservoir. The framework was evaluated on a simulated Maari subsurface hydrogen storage and exhibited strong classification performance and prediction of estimated hydrogen recovery from the subsurface environment.