Blending Scheduling And Operations Optimization For Refining Process

Refinery production and processing of non-ferrous metal refining are typical of refining process.The refining process usually includes the blending of material flow,physical or chemical processing reactions,material storage,etc.Scheduling of material flow and operation optimization in refining process can improve production efficiency,reduce the consumption of raw materials,improve the energy efficiency and increase the production profits.Therefore,it has great theoretical and practical value to study the blending of material and operation optimization of typical refining process.In this thesis,two problems are mainly focused on.One is the blending scheduling of material in refining front-end,and the other one is operation optimization problem refining process.A generalized disjunctive programming model for blending scheduling of material problem in refining front-end is established,and an improved logic-based outer approximation algorithm is presented to solve the problem.Using Kriging algorithm based on estimation of distribution of metal smelting process quality and production and operation condition analysis and optimization.The Kriging algorithm based on distribution estimation is used to analyze and optimize the metal grade and production operating condition in refining process.In addition,the system of material balance and operation optimization is designed and developed.Details are as follows:1)The GDP models with linear disjunctions are reviewed.Two reformulation methods for GDP models including Big-M and convex hull reformulation are compared by their advantages and disadvantages through numerical experiments.The outer approximation algorithm based on logic is implemented and extended to solve the generalized disjunction programming model,and the difference with outer approximation is compared.2)A generalized disjunctive programming model based on continuous time is established for blending scheduling of material problem in refining front-end.According to the characteristics of model,an improved logic-based outer approximation algorithm is designed to solve the model and compared with the Big-M transform method and LOGMIP software.Experimental results show that the improved logic-based outer approximation algorithm outperforms the existing methods in solving time and iterations.3)A Kriging algorithm based on distribution estimation is proposed to solve the optimization problem of germanium metal and operating condition in refining process.The distribution estimation method is used to solve the fitting of the variation function in Kriging algorithm,and the weight coefficient of Kriging is optimized to achieve the purpose of analyzing and optimizing the grade of germanium metal and operating condition.The production operation condition obtained by the analysis and optimization can provide the reference for enterprise to arrange the production.4)The system of refining process material balance operation optimization is developed,which is already applied to the actual production enterprises and achieve the enterprise's digital platform management.Based on the production data,analyze the reason of the losses of germanium metal materials.Then according to germanium grade,introduce the algorithm of embedded system to optimise the condition of production operation.