| Energy consumption is the key to restrict the development of cement industry,cement grinding system is the key production process of cement production,optimization of the key parameters of power consumption in cement grinding process can provide a basis for energy saving and consumption reduction in cement production process.Subject to power consumption,energy consumption of cement grinding process of cement grinding process is a time-varying,continuous,and the process of mutual coupling between each process variable,power consumption in cement grinding process is vulnerable to operation of the equipment,raw materials fluctuation and the change of the process variables,the influence of such factors by analyzing the mechanism of cement mill run hard to establish accurate mathematical model to control the power consumption.The purpose of this paper is to establish the power consumption prediction model of cement mill production by data-driven method,and modify the production process variables to adapt to the new working conditions by constantly updating the prediction results.Based on the prediction model,the optimization model of cement grinding process variables was built by combining the constraints of cement grinding process equipment and power consumption index.DC-BAS algorithm was used to optimize and solve the process variables,so as to realize the optimization of cement grinding production process based on power consumption prediction.The specific research work is as follows:(1)Based on the process mechanism of cement grinding process,the related equipment influencing the power consumption of cement grinding process was analyzed,and the related variables influencing the power consumption of cement grinding process were further selected.The mutual information method is used to measure the correlation coefficient between power consumption related variables and power consumption.The current manual empirical analysis is converted to the selection based on the correlation coefficient,and the key variables affecting the power consumption in the cement grinding process are screened out to reduce the data redundancy,which lays a foundation for the establishment of the power consumption prediction model in the subsequent cement grinding process.(2)Aiming at the characteristic problems of cement grinding process,such as time-varying ductility,non-linearity and uncertainty,and common problems such as immediacy of production requirements.Make full use of the variable data and temporal characteristics of the target variable,according to the delay time of cement grinding process to determine the length of time window to arrange all the variable data as input layer,using Improved Particle Swarm Optimization(IPSO)algorithm for High-Performance Extreme Learning Machines(HP-ELMs)of structure parameter Optimization,The influence of variable time-varying delay characteristics on power consumption prediction accuracy is eliminated,and the power consumption prediction is realized.(3)The power consumption prediction model was taken as the objective function,combined with the equipment operation and working condition information in the cement grinding process,and the Dynamic Correction Beetle Antennae Search(DC-BAS)algorithm was used to solve the objective function,and the optimal process variable suitable for the current working condition was obtained.The experimental results show that the power consumption prediction model established in this paper has high prediction accuracy,and the power consumption can be effectively reduced when the optimized process variables are substituted into the objective function as set values. |
