Oil and Gas Facility Layout Optimization — A New Clustering Based Approach

Abstract Field development planning involves substantial investment; therefore, it must be carefully optimized to ensure cost-effective and efficient operations. A key component of this process is facility placement, which, when combined with well placement and trajectory design, becomes a highly complex optimization problem, particularly during the early concept-screening phase. At this stage, multiple parameters and numerous development scenarios need to be evaluated within a limited timeframe, creating the need for an integrated optimization workflow that minimizes cost while maximizing overall field performance. Previous studies have presented various optimization methods that either require exhaustive search, making them prohibitively slow, or require less computational time at the expense of optimality. These methods often struggle to capture complexities and locations limitations of real fields. In this work, a novel clustering-based algorithm is developed to optimize facility placement, which can employ any of the following clustering methods: K-means, hierarchical clustering, and Gaussian mixture models (GMM). The algorithm accounts for real fields’ complexities and optimizes multiple layers of facilities. The developed approach is benchmarked against a metaheuristic algorithm, with comparisons drawn across multiple test scenarios with various levels of complexity, including multilayer facility layouts, prohibited areas, and topological complexity. Results show that the developed clustering-based algorithm is modular, robust, and performs substantially better than the metaheuristic approach in terms of optimality, but at the expense of computational requirements.