Abstract
Carbon capture and storage (CCS) projects typically involve a diverse arrayof stakeholders or players from public, private, and regulatory sectors, eachwith different objectives and responsibilities. Given the complexity, scale,and long-term nature of CCS operations, determining whether individualstakeholders can independently maximize their interests or whethercollaborative coalition agreements are needed remains a central question foreffective CCS project planning and management. CCS projects are oftenimplemented in geologically connected sites, where shared geological featuressuch as pressure space and reservoir pore capacity can lead to competitivebehavior among stakeholders. Furthermore, CO2 storage sites are often locatedin geologically mature basins that previously served as sites for hydrocarbonextraction or wastewater disposal in order to leverage existinginfrastructures, which makes unilateral optimization even more complicated andunrealistic. In this work, we propose a paradigm based on Markov games to quantitativelyinvestigate how different coalition structures affect the goals ofstakeholders. We frame this multi-stakeholder multi-site problem as amulti-agent reinforcement learning problem with safety constraints. Ourapproach enables agents to learn optimal strategies while compliant with safetyregulations. We present an example where multiple operators are injecting CO2into their respective project areas in a geologically connected basin. Toaddress the high computational cost of repeated simulations of high-fidelitymodels, a previously developed surrogate model based on the Embed-to-Control(E2C) framework is employed. Our results demonstrate the effectiveness of theproposed framework in addressing optimal management of CO2 storage whenmultiple stakeholders with various objectives and goals are involved.