Abstract: To address the challenges of numerous and diverse equipment, coupled hydraulic and thermal characteristics, mutually constrained operating parameters, and the difficulty in comprehensively optimizing operation schemes for large-scale crude oil depot systems, this study constructs a mixed-integer nonlinear programming multi-constraint operation optimization model with the objectives of minimizing total energy consumption, total operational cost, and total carbon emissions. Based on the characteristics of the model, an improved NSGA-III algorithm is proposed for solving this multi-objective optimization problem. The improved algorithm introduces heuristic initialization and a dynamic penalty function, and integrates a Memetic framework based on sequential quadratic programming, thereby accelerating convergence speed and enabling fine-grained local search over discrete and continuous variables. The results show that the hypervolume and inverted generational distance metrics of the solution sets obtained by the improved algorithm are significantly superior to those obtained by traditional NSGA-II, NSGA-III, MOEA/D, and RVEA algorithms. The optimized operation scheme reduces total energy consumption, total operational cost, and total carbon emissions by 10.07%, 14.51%, and 7.41%, respectively. The optimal crude oil throughput ranges from 10000 to 16000 t/d. Ambient temperature significantly influences the start–stop schedule and heat load of heaters; therefore, the system should adopt a seasonal heating strategy of “winter operation and summer shutdown,” achieving energy conservation, cost reduction, and carbon emission reduction while maintaining safe and efficient operation.

Key words: oil depot process system, equipment coordinated operation, multi-objective, operation optimization, improved NSGA-III algorithm