Optimization Of Comprehensive Index For Cement Combined Grinding Process

Cement grinding process is a key part in cement production,which directly affects the quality,output and other technical and economic indicators of cement products.With the gradual deepening of people's understanding of the grinding process,cement grinding technology has shown a diversified trend.Throughout the current cement grinding process,the combined grinding system consisting of “roller press + ball mill” shows the advantages of “crushing instead of grinding” and “more crushing and less grinding”,which is widely used in industry.The primary task of grinding is to control the particle size index of the cement product within the target value range specified by the process,while ensuring that the mill load is at the optimal working point and improving the grinding efficiency.Therefore,it is important to study how to optimize the index of cement combined grinding process to improve the quality of cement product.Focusing on the problem,the main research contents of this paper are as follows:(1)Aiming at the complexity of the production process of cement combined grinding process,through the in-depth analysis of the dynamic characteristics and manual operation mode of the combined grinding process,this paper proposes that the pre-grinding of the roller press and the final grinding of the ball mill should be taken as a whole to study and establish a model which can show the changing law of the whole process.According to the analysis and selection of operation variables,state variables and process indexes in the process of cement combined grinding,a combined model combining mill load model and cement particle size model was established by radial basis function(RBF)neural network.By using the actual production data of a cement plant to simulate and verify,and comparing with the simulation results of back propagation(BP)neural network,the results show that RBF neural network has higher fitting ability and smaller error.Therefore,the combined model based on RBF neural network is reasonable and effective.(2)Aiming at the large fluctuation of particle size in grinding process caused by setting value of control loop depending on operator's experience,the strategy and method of intelligent optimization of cement combined grinding process were proposed.The optimization method consists of a pre-set module based on case-based reasoning,a cement granularity prediction module based on RBF neural network,and a feedforward compensation module and feedback compensation module based on rule-based reasoning.The optimal set value of the control loop is provided by the target value of the indicator and the boundary condition,thereby ensuring that the actual value of the cement quality index is within the target value range.In turn,the purpose of increasing production and reducing energy consumption is achieved.(3)Based on the above intelligent optimization method,an intelligent optimization system for cement combined grinding process was designed.The optimization system can not only give reasonable and effective setting values under the current working conditions according to the target value of granularity and boundary conditions,but also compensate for the pre-set values in time according to the operation effect,which can effectively reduce the occurrence of quality disqualification caused by improper operation of operators.Based on VB programming language,an intelligent optimization setting software was developed and applied to cement grinding production site.The application shows that the software has excellent performance and has great practical significance for guiding cement grinding production.