Soft Computing Tools for Multiobjective Optimization of Offshore Crude Oil and Gas Separation Plant for the Best Operational Condition

The selection of operating conditions in the oil and gas separation plants is obtained through data monitoring or through the experience of the operating personnel which can bring operational difficulties leading to inefficiencies and capital loss. The modern techniques of soft computing including artificial intelligence and genetic algorithm-based optimization can add value to this operation. In this work, five key controllable design variables of an oil and gas separation plant are optimized considering two performance indicators (oil flow productivity and gas compression power). The physical model of the plant is simulated using ASPEN HYSYS, and a digital twin model is generated using an artificial neural network. It is followed by a multiobjective optimization using non-dominating sorting genetic algorithm II to obtain the Pareto front. The results have indicated that operational optimization can enhance oil production by up to ∼6.2% and decrease the compression work by ∼3.2%. It is concluded that the proposed operation of the plant is energy-efficient and can increase productivity.