Integrated Optimization And Prediction For Enterprise-wide Production Of Refining Process

The enterprise-wide production in refining process has certain characteristics,such as large amount of production equipments,complicated technical process,high temperature,large batch production and so on.These characteristics determine the high energy consumption and the high pollution in its production process.With the increasing concern of the energy problem and the environmental problem,how to improve the enterprise's production management ability,improve the production efficiency,reduce the production energy consumption and improve the enterprise's profit under the current resource condition is an urgent problem..The enterprise-wide production in refining process problems include the whole process of production planning and scheduling problems.Production planning is used to determine the amount of production.Production scheduling is used to make specific production schedules and complete production planning tasks.Taking into account the continuous separation of refining characteristics,the production yield of the production process directly affects the production of the final product.Therefore,it is possible to predict the product yield of the refining process to more accurately control the actual production of the product.The thesis focuses on integrated optimization and prediction for enterprise-wide production of refining process,with the refinery and metal smelting process as the representative.Based on the actual research,the refinery production process,energy consumption and production management demand are analyzed,and the production plan and production scheduling model of the whole process of the refinery are given.Based on data analysis of the correlation between metal grade and production conditions in refining process,the product recovery rate of the refining process is predicted.Finally,a decision support system for smelting process is developed.The details are as follows:1)Aiming at the requirements of refinery production plan,this thesis analyzes the whole production process of refinery with energy consumption considered,abstracts the whole production plan,and establishes the model of refinery production planning with energy consumption considered.Based on the actual production data of a refinery,by comparing with the Aspen PIMS and refinery process model in the published literature,the numerical results show that the proposed model can obtain the high precision plan.2)Based on the production plan of the refinery,the problem of refinery production scheduling with energy consumption considered is proposed,and then the continuous time mathematical model is established,which is validated by numerical experiments.Finally,the refinery planning and the scheduling program are integrated,and the production management program of the whole refinery process with energy consumption considered is obtained.3)From the metal refining process to obtain product yield and operating conditions of the actual production data.the Kriging algorithm based on least squares support vector machine(LS-SVM)is used to predict the product recovery rate in the refining process through data analysis,the experimental results show that the prediction error of product yield is less than 6%.4)The material balance decision support system of smelting process is designed and developed.The system realizes the enterprise real-time management to the production process data,and an improved prediction algorithm for product yield is built by embedding,to achieve a refining process for a given part of the cycle of product yield prediction.