Research On Scheduling For Mixing And Logistics Operations In Refinery

The mixing and logistics scheduling of refinery oils includes the mixing and logistics scheduling for crude oil and the on-line blending scheduling for product oil,which determines the safety of production process,the quality of oil products and the economic benefits in refinery.The mixing and logistics scheduling for crude oil is the input source of raw materials for refinery production.Considering operation safety,optimizing crude oil mixing and logistics operation is of great significance for reducing the storage and transportation costs of crude oil,ensuring the continuous supply of refinery raw materials,and for the safe and stable operations of refinery production process.The on-line blending scheduling for product oil is important for the process of gasoline production in refinery.Optimized blending recipes and scheduling schemes can further improve the quality of product oil while satisfying the quality standard,which is also of positive significance to reduce production costs and resource consumption of refinery,and reduce the environmental pollution caused by the use of gasoline.Based on the characteristics of safe production process of refinery,this thesis addresses the crude oil mixing and logistics scheduling problem and the on-line blending scheduling for product oil in refinery.We formulate two separate generalized disjunction programming models and design an improved global logic-based outer approximation algorithm.Finally,we design and develop a decision support system.The details of this thesis are as follows:1)Considering operation safety,a generalized disjunction programming model for crude oil mixing and logistics scheduling is formulated,based on the continuous time representation.An improved global logic-based outer approximation algorithm is designed to solve the model.Compared with the LOGMIP solver,the proposed algorithm improves the solving precision and accelerates the convergence process.2)Combined with the uncertainty of the arrival time of vessels,the problem of crude oil mixing and logistics scheduling with consideration of uncertainty is defined.A generalized disjunctive programming model based on scenario tree is formulated.Based on the characteristics of the proposed model,a logic-based outer approximation algorithm is designed and implemented to obtain optimized crude oil scheduling schemes under different scenarios.3)The non-convex generalized disjunction programming model for the on-line blending scheduling problem is improved,and the improved global logic-based outer approximation algorithm is used to solve the model.Optimized blending recipes and scheduling schemes for the on-line blending scheduling are obtained,which can further improve the quality of gasoline.4)A decision support system for oils mixing and logistics scheduling in refinery is designed and developed,the decision support subsystem for crude oil mixing and logistics scheduling is mainly designed and developed.The decision support system can realize effective management of related data information in the process of scheduling operations,and provide optimized scheduling schemes by using the embedded optimized model and algorithm.