earticle

영어

The optimal scheduling of crude-oil operation in refineries has been researched by various groups during the past decade. Different mixed integer linear programming or mixed nonlinear programming formulations are derived. This paper proposes a novel MINLP formulation with multiple charging tanks charging a distiller overlapingly and oil residency time constraint that is based on multi-operation sequencing with synchronized start time (MOS-SST). It is different from previous formulations because of its consideration for oil residency time constraint and multiple charging tanks charging a distiller overlapingly. A branch and bound algorithm combined with the Lagrangian Relaxation Subgradient Optimization algorithm is proposed to solve this model, resulting in a satisfactory optimal result.