Intelligent Optimization of Water Alternating Gas Flooding with Carbon Dioxide: A Win-Win Strategy for Enhancing Oil Recovery and Carbon Storage

Water alternating gas (WAG) flooding with carbon dioxide (CO2) is one of the most commercially viable technologies for enhancing oil recovery and CO2 storage. This study proposes an efficient surrogate-assisted optimization method based on the active learning strategy and surrogate ensemble for multi-objective WAG well control parameters optimization named ALSA-MOPO. In the ALSA-MOPO method, three frequently used surrogate models, the radial basis function network (RBFN), Gaussian process regression (GPR), and support vector regression (SVR) are applied to construct the surrogate ensemble. In addition, an active learning strategy is adopted to reduce the sample of establishing surrogate model and query for the worst and best samples based on the surrogate ensemble using particle swarm optimization (PSO) which is infilled into the training dataset for improving the quality of the surrogate ensemble. Furthermore, the ZCD which is a complex fault block reservoir located in the North China Basin is adopted to test the effectiveness and practicability of the ALSA-MOPO method. The optimization results demonstrated that the ALSA-MOPO method was about thirty times faster than the numerical simulation-based methods. More importantly, the method presented in the research can provide oil companies and researchers with a reference for oil production and CO2 storage. 1. INTRODUCTION In the dynamic landscape of global energy demand and environmental sustainability, the oil and gas industry towards significant challenges of maximizing hydrocarbon recovery from underground reservoirs while mitigating the environmental impact of fossil fuel extraction and utilization (Bajpai et al., 2022). As the world transitions towards a low-carbon future, there is an increasing emphasis on innovative oil production strategies that not only enhance oil recovery efficiency but also facilitate the storage of carbon dioxide (CO2) underground (Krevor et al., 2023). In this context, water alternating gas (WAG) as a win-win scenario and one of the most effective ways to enhance oil recovery (EOR) while storage of CO2 has drawn widespread attention in recent years. It involves injecting water and CO2 gas alternately into the reservoir to improve oil recovery and assist storage of CO2 efficiency. Meanwhile, the WAG flooding with CO2 has emerged as a prominent EOR technique, which exploits the synergistic properties of water and CO2 to displace oil within reservoirs (Wang et al., 2023). Specifically, the alternating injection of water and CO2 serves multiple mechanisms: water serves as a carrier fluid to facilitate the mobilization of oil, while CO2 serves a dual role by reducing the viscosity of the oil and exerting pressure to drive oil toward production wells. Furthermore, the injection of CO2 offers the extra benefit of carbon storage, as the greenhouse gas is sequestered underground, thereby mitigating its release into the atmosphere (Saraf et al., 2023). Under these situations, searching for optimal well control schemes for the WAG flooding process will maximize development benefits and the amount of CO2 storage into the reservoir.