Research On Multi-objective Optimization Model And Robust Optimization Algorithm For Cement Firing System

The cement firing system is one of the core equipment of cement production,and the raw meal is calcined in the rotary kiln to generate cement clinker,which consumes a large amount of energy.Optimizing the variables associated with the cement firing system can help save energy and reduce consumption in the cement production process.At present,China’s cement manufacturing industry is still facing problems such as high energy consumption and poor product quality.The main difficulties are that there are many equipment and parameters of cement firing system,complex production mechanism,it is difficult to accurately establish the mechanism model of cement firing system,and due to the dynamic change of working conditions and process parameters,it is difficult to ensure that the optimization results are effective for a long time.This paper aims to establish a coal consumption and quality model of cement firing system in a data-driven manner.The objective function is built based on the model,and the optimization model of cement firing system is constructed by combining the equipment constraints and index constraints of the cement firing process.The robust multiobjective optimization algorithm is used to obtain an effective optimization solution to ensure the stable quality of cement clinker and reduce coal consumption.The specific research content is as follows:(1)Aiming at the optimization of coal consumption and production indicators of cement firing system,a multi-objective optimization model of cement firing system is established with reducing coal consumption and ensuring quality as the optimization goals,key variables as decision-making variables,and comprehensive consideration of production indicators,equipment capacity,variable constraints and other factors.At the same time,aiming at the problems of time-varying delay,nonlinearity and large time scale difference between decision variables and production indicators in the clinker production process,a convolutional neural network based on time series and convolution-cycle high-speed network are used to predict coal consumption and quality,which solves the problem that it is difficult to establish a cement firing system model.(2)Aiming at the problem of deterioration of optimization performance caused by the fluctuation of decision variables,a robust multi-objective differential evolution algorithm for cement firing system is proposed.In the process of algorithm iteration,the robustness of individuals is calculated,and a robust evolutionary strategy is designed,and the population selects suitable candidate solutions during the iteration process to ensure that the robust optimal solution is searched.The algorithm solves the problem of how to ensure that the optimization results are effective,so as to achieve the goal of reducing energy consumption and ensuring clinker quality.(3)The clinker production data of a cement enterprise was used to conduct experiments to verify the proposed model and algorithm.The model accuracy of convolutional neural network and convolution-loop high-speed network based on time series was analyzed.By comparing the optimization experimental results with the historical data,the performance of the proposed optimization method is proved.Through the experimental comparison with the traditional multi-objective optimization algorithm,the advantages of the proposed optimization algorithm in reducing coal consumption under the condition of fluctuation of decision variables are proved.