Research On Load Forecasting Algorithm Of Cement Mill

Cement,as a basic material for the production of concrete and mortar,is widely used in construction,water conservancy,transportation and other engineering fields.After China's reform and opening up,the energy industry has developed rapidly,accompanied by increasing environmental pollution problems,Emission of a large amount of harmful gases,any improvement in a process or equipment will promote the entire production process,achieve energy conservation and emissions reduction,and promote the sustainable development of the cement industry.This article takes a 4000t/d cement plant in Ankang City,Shannxi Province as the object,and on the basis of fully analyzing the new dry process cement production process,it summarizes the research results at home and abroad.According to the characteristics of the large span,wide range of cement production line,and many equipment involved,Zhejiang Central Control WebField ECS-700 distributed control system.And according to the problem that the mill load is difficult to detect during the grinding process,different mill load prediction models are proposed.The main research contents of this paper are as follows:(1)Completed the overall control architecture including overall system design,hardware configuration,and software configuration according to control requirements and production processes.The main process of the new dry cement production process is divided into four sections of raw material preparation and grinding,preheater thermal decomposition,clinker calcination and cement making.Each section is provided with a field control station.The communication network between the operation nodes adopts industrial Ethernet based on TCP/IP,and the control network adopts redundant optical fiber ring network,which solves the problem of long-distance communication with stable operation of the production line.Use the VFExplorer control platform forhardware configuration,calculate control points according to production line equipment control and transmitter signal access requirements,count I/O points and leave appropriate margins,determine the number,model and related configuration of the controller and I/O modules.After the hardware configuration is completed,the VFFBDBuilder programming software is used to build a chain program according to the equipment control requirements.Finally,the VFHMICfg software is used to configure the upper computer screen and link the site data according to the process flow.(2)Most of the energy consumption in the cement production process is used to grind cement raw materials,and the mill load is an important indicator for evaluating the operating state of the mill,so it is particularly important to be able to accurately judge the mill load state.In order to respond to the national call for sustainable development,reduce production energy consumption,and aim at the problem that the load is difficult to detect during the grinding process of the mill,this paper proposes a mill load prediction model based on an improved particle swarm optimization algorithm RBF neural network.Based on the RBF prediction model,the PSO algorithm is used to optimize the center vector of the RBF network,the base width parameter and the connection weight of the hidden layer to the output layer.By improving the inertia weight factor,a non-linear inertia weight decreasing strategy is proposed to balance the local and The contradiction between the global particle search capabilities makes it possible to find the optimal solution quickly and accurately.The simulation results show that the predicted value of the RBF model deviates greatly from the actual value.The prediction accuracy of the PSO-RBF and IPSO-RBF models are much higher than the RBF.The predicted value of the PSO-RBF model is close to the actual value.The predicted value of the IPSO-RBF model The change curve is almost consistent with the actual value,with the smallest error.Compared with PSO-RBF,the decision coefficient of the IPSO-RBF model is increased by 0.0795,and the root mean square error RMSE,mean absolute error MAE,and mean square error MSE are respectively reduced by50.1%,48.1% and 75.1%,fully confirm the effectiveness of the improved algorithm.In this paper,a DCS system for a cement plant is designed according to the functional requirements of the production line.Aiming at the problem of difficult to detect the mill load,comprehensively consider the influencing factors of the mill load,establish a relevant prediction model of the mill load,and combine the engineering background with theoretical research based on Follow-up related research provides technical support.