Research On Prediction Of Vertical Roller Mill Operation Based On IGM And OKELM

As a new kind of high efficient and energy-saving grinding equipment,it is the key equipment in cement production,and its long-term and reliable operation are very important to the benefit of the enterprise.In this paper,the prediction of the running state of the pin parts of the vertical roller mill is studied,the research results can lay a foundation for the predictive operation and maintenance of the vertical roller mill.It is of great theoretical and practical significance to improve the reliability,safety and maintainability of such complex equipment such as vertical roller mill.In this paper,two kinds of improved single prediction methods are proposed based on the error minimization of the running state prediction problem of the vertical roller mill pin parts,which are the improved grey prediction method and the optimized kernel extreme learning machine method.On the basis of this,in order to synthesize the advantages of the two kinds of single prediction methods and obtain better prediction results,the improved grey prediction method is used to predict the original data sequence,the results of the improved grey prediction model are subtracted from the original data,and the residual sequence is obtained.The optimized kernel extreme learning machine method is used to fit the residual sequence to correct the prediction results of the improved grey prediction method,and the combined forecasting algorithm is obtained to realize the prediction of the running state of the vertical roller mill.(1)Research on the running state of vertical mill based on the improved grey prediction method.In order to solve the problem of fitting error of the model initial value of grey prediction model(GM(1,1)),a correction coefficient is used to reduce the fitting error of the model initial value.In addition,in order to improve the accuracy of the method,the quantum genetic algorithm is used to optimize the model parameters of the method.Based on this,an improved grey prediction model is established and applied to the prediction of the running state of the vertical roller mill pin parts.The experimental results show that the improved grey prediction model can increase the accuracy of the algorithm to a certain extend and obtain better prediction effect.(2)Research on the method of vertical mill running state based on the optimized kernel extreme learning machine.In order to improve the prediction accuracy of the model,the artificial bee colony algorithm is used to optimize the model parameters of the kernel learning machine.In addition,due to the influence of the high-dimensional kernel matrix in the kernel extreme learning machine,which leads to the increasements of the complexity of the output matrix calculation and training time,the ImprovedNystrom matrix decomposition method is used to reduce the time complexity of the model training.The prediction experimental results of the stress state of the vertical roller mill pin parts show that the optimized kernel extreme learning machine can improve the precision of the algorithm and reduce the time complexity of the algorithm.(3)The prediction of the running state of vertical mill by the combination forecasting method of residual compensation.Aiming at the two improved prediction models proposed above,the improved grey prediction method is used to predict the principal component information of the data sequence,and the residual sequence of the improved grey prediction model is fitted by the optimized kernel extreme learning machine to reflect the fluctuation characteristics of the data sequence,the residual prediction results are compensated to the preliminary prediction results of the improved grey prediction model to obtain the final prediction results.We compare the residual forecasting method based on the residual error compensation with other prediction methods.It can be seen that the combination forecasting method has higher prediction precision and fitting effect,and it is suitable for the actual prediction of the prediction of running state of the vertical mill.Based on the above research,the vertical mill running condition prediction system is designed and developed.