Robust Engineering Strategy for Scheduling Optimization of Refinery Fuel Gas System

As a byproduct of the oil refining process, fuel gas is the primary energy source of refineries. Considering self-generated and purchased fuel gas simultaneously in an optimization model will cut down energy cost and reduce carbon emissions in oil refineries. A mixed-integer linear program (MILP) has been built in our previous work. However, due to the fluctuation in the fuel gas generation and consumption, theoretical scheduling solutions may become infeasible or inaccurate. This article presents a robust engineering strategy for validating the model to variable conditions in four aspects: model precision, solving performance, optimization effect, and execution. The proposed strategy has been applied to a fuel gas system in one of the largest oil refineries (LRF) in China to ensure model feasibility, necessity, and effectiveness. The implementation results show that the proposed method reduces costs up to 5.63% through the single-period operational optimization and up to 7.76% in the multiperiod scheduling.