面向MES的炼油厂调度建模策略研究

Strategy Study on MES-oriented Scheduling Model in the Refinery

基于建模平台PS(Production Scheduler),以炼油厂作为研究对象,以效益类和绩效类目标结合的多目标规划建立润滑油调度模型,并用分支定界法、线性规划方法进行求解;在包含性质约束的汽油调和模型中,以路径层和性质层分解的策略分别建立混合线性规划模型和不包含整型变量的非线性规划模型,在性质层模型求解中运用SLP(连续线性规划)方法求解。建模、求解和仿真结果证明,以线性规划为基础、对象化方法为思路、结合分层次的分解策略建立离散时间调度模型的多目标规划方法,适用于建立炼油厂各关注重点的局部模型和全厂模型,结合分支定界和SLP的线性规划求解策略,适用于大规模调度模型的求解。

Based on the model-forming platform of PS(Production Scheduler),using the refinery as the object of study,the lubricating oil scheduling model was established by multi-objective programming combining efficiency object with performance object.B&B(Branch and Bound)and LP(Linear Programming)algorithm were adopted to solve the model.In the quality-constrained gasoline blending model,MILP(Mix-Integer Linear Programming)model and NLP(Nonlinear Programming)model without integer variable were established respectively by the decomposition strategy of route and quality.The SLP(Successive Linear Programming)method was adopted to solve the quality model.Based on the Linear Programming and object-oriented method,combining classified decomposition strategy,discrete time scheduling model was established.The results of modeling,solving,and emulation proved that the multi-objective planning method was applied to partial model and integral model in the refinery.The linear planning solution strategy combining B&B and SLP algorithm was applied to solving large scale scheduling model.