Unlocking Potential: Automated Optimization of Water Injection in Giant Gas Cap Reservoir for Peak Performance and Economic Gain

Abstract In the dynamic landscape of oil and gas field management, the growing emphasis on automation, AI, and ML technologies is evident. This study introduces an innovative automated workflow tailored for the complexities of vast offshore carbonate reservoirs hosting gas caps, with a specific focus on optimizing water injection field development strategies. Utilizing commercial applications and Python scripting, the research meticulously analyses economic factors to refine the development plan. This holistic approach aims to achieve a delicate balance between maximizing profitability and minimizing costs, thereby paving the way for heightened operational efficiency and sustainable growth. We present an innovative workflow tailored for optimizing the field development plan (FDP) of one of the largest and most intricate offshore gas cap reservoirs located in the UAE. Overcoming various challenges, including (1) determining the optimal number of inner ring water injection wells to maintain gas cap integrity, (2) addressing communication issues between multiple reservoirs leading to improper balance between water and gas injection, (3) automating designing and planning of well trajectories, and (4) integrating economic evaluations for optimization. The methodology centered on an engineered FDP optimization and automation workflow. This involved varying control parameters and executing integrated workflow of geological well trajectory planning, dynamic simulation, and Python-scripted workflows to generate multiple realizations based on these parameters. Subsequently, the scenarios were simulated, and the FDP was automated considering economic factors. The approach starts with geological workflow that automates well planning, trajectory designs and can generate multiple realizations varying number and location of inner ring water injectors. Each well-planned realization is automatically incorporated and submitted into dynamic simulation followed by economic evaluation. The whole workflow is orchestrated using commercial optimization and uncertainty assessment software. The first batch of realizations samples the multi-dimensional optimization solution space and used to build the initial set of multi-variate proxy models. Proxy models are used later to submit successive optimization batches to maximize the economical objective function, where proxy models automatically get updated and improved by submitting more and more batches. The implementation of the automated FDP optimization workflow yielded promising results. We achieved a 17% reduction in well count while ensuring reservoir pressure maintenance and gas cap containment. Moreover, by automating the workflow, we accelerated the FDP optimization process by 60%. Furthermore, there is potential for additional cost optimization by expanding the workflow to optimize the number and placement of future producers.