Research On Simultaneous Cyclic Scheduling And Optimization Poblems Of The Refinery CSTR Processing

Refining enterprises have the characteristics of capital and technology intensive, continuous production processes, large volume products and broad market. The type of ends of the supply chain in the refinery is the discrete and intermediate is continuous. In general, with the transport and marketing constraints and resulting in inventories of raw materials and products, the entire production process has a combination of discrete and continuous characteristics. Based on the characteristics of the refinery production scheduling control, production and management decision-making in the enterprise has played a decisive role. For the production planning and scheduling control study, the main contents of this thesis are outlined as following:(1) Based on the background of refinery production scheduling and optimization problem, it has outlined the refineries business processes, analyzed refinery CSTR (Continuous Stirred Tank Reactor) production and scheduling process optimization principle, classification, characteristics of the problem.(2) Provided an important issue of multi-product production cyclic scheduling and optimization at the refinery, analyzed the constraints of the problem. Based on the key elements of the production process, the simultaneous scheduling and optimization problem is casted as a Mixed-Integer Dynamic Optimization (MIDO) model. Based on orthogonal collocation on finite elements, it is used to transform into a Mixed-Integer Nonlinear Programming (MINLP) model.(3) It has systematically introduced tabu search algorithm and further researched to improve the intelligence optimization method. The hybrid method combines the tabu search local search and a strong expansion of the scope of neighborhood search capabilities in order to better improve the performance of the algorithm.(4) Based on analysis methods of solving problem, it has used the finite element orthogonal collocation method and feasible path method to decompose the model, briefly sketched the flow chart of solving model. Comparison with the tabu search algorithm, the improved tabu search algorithm, and ILOG CPLEX mathematical programming algorithm, the results have been analyzed to verify the validity of the model and algorithm.Finally, the conclusion and perspectives are given in the end of the thesis, pointing out that the next phase of the study and research.