Modeling And Simulation For Oil Refining Process Scheduling

Under globalization and intense competition in the market, efficient production scheduling is of the important guarantee to help oil refineries improve competitiveness and gain more profits. However, due to the complexity of the refining processes and the production scheduling problem NP-Hard nature, although the theory research has made some progress, yet in practice is still mainly based on human experience。A programming model with reasonable size for scheduling is not able to describe in detail the production processes and often the obtained solution may not have operational flexibility, thus " Simulation+Heuristic rules" is considered as a future trend of refinery scheduling. Whereas the research specifically for refinery scheduling simulation is very limited. Some related work is carried out in this article after the major issues on refinery scheduling needs of simulation, refining process modeling and simulation,and VV&A are reviewed, as well as a software prototype for refinery scheduling simulation is developed. The main contents of this thesis are as follows:1. Firstly, a descriptive model based simulation framework is proposed. After that, a MES layer descriptive model is defined and discussed base on Enterprise-Control system Integration standard and multiple resolution modeling theory, and its E-R conceptual data model is designed. At last the aggregation and disaggregation relationship of descriptive model among PCS/MES/ERP is discussed.2. Statistical modeling of refining process unit is discussed considering factors like material-flow, key properties, unit energy consumption and production uncertainty. And the sorted and improved corresponding statistical models are introduced into scheduling simulation as an attempt. Sulfur content distribution model and unit power consumption prediction model are presented along with material-flow model as an instance.3. According to refinery scheduling demands for simulation, a visual simulation system is built making use of database technology and object-oriented technology. In this system a hybrid simulation strategy is applied taking advantages of both steady state process simulation and discrete event system dynamic simulation. Meanwhile, a graphical descriptive model base and a process unit simulation model library are developed to assist user in flowsheeting and simulation.4. A real domestic refinery is taken as a case study object to validate modeling and simulation methods proposed in this article. The main processes include flowsheet validation, unit simulation model validation and multi-batch refinery-wide steady material-flow simulation validation. The results of validation demonstrate the feasibility and effectiveness.Finally, a summary of this thesis is concluded and further research objectives and directions are pointed out.