Prediction Of Coal And Gas Outburst Grade Based On Optimized Kernel Extreme Learning Machine

Coal and gas outburst is a highly destructive geological disaster occurred in the process of coal mining,The occurrence of outburst disaster will seriously threaten the safety of underground workers and even have a significant impact on the national economy.Therefore,it is of great significance to accurately predict coal and gas outburst dangerousness and guarantee the safety of mine production.The content of this article of this article is mainly explained from the following aspect:(1)In order to solve the problem that the high-dimensional and tedious data of outburst factors affect the calculation speed and prediction accuracy of the model,the Laplacian Eigenmaps algorithm(LE)was used to extract the features of the outburst factors and remove the redundant data.In order to verify its effectiveness,LE algorithm was compared with principal component analysis,kernel principal component analysis and local linear embedding to verify the superiority of LE algorithm.(2)Aiming at the artificial blind selection of normalization parameters and kernel function parameters during KELM network training leads to defects in poor model classification.The relevant parameters of KELM were optimized combined with CSA algorithm to improve the classification accuracy.In order to avoid the shortage of local optimality in the iteration process of the traditional CSA algorithm,multiple strategies including tent chaotic sequence,adaptive perception probability,adaptive regulation step length and Brownian motion are introduced to improve the CSA algorithm,and compared with PSO,GOA and CSA to verify the superiority of ICSA algorithm.Finally,the data after dimension reduction of LE algorithm is used as input of model,and establish the coal and gas outburst model of KELM optimized by ICSA algorithm.(3)To further illustrate the superior performance of ICSA-KELM coal and gas outburst model.Firstly,it is verified that the accuracy of KELM model after feature extraction of LE algorithm was improved that without feature extraction.Then,the KELM model was compared with ELM,SVM and RVM models by using feature extraction data as the input of four models.Next,the analysis of ICSA-KELM model with CSA-KELM model,PSO-KELM model and GOA-KELM model shows that this model can effectively improve the prediction accuracy of coal and gas outburst.Finally,the convergence time of the four algorithms was compared,and the ICSA-KELM model can reach the convergence state quickly and have high accuracy.Therefore,the advantages of the model proposed in this paper are verified from the perspectives of prediction accuracy and convergence time,which can provide support for the prevention and control of coal and gas outburst disasters.The thesis contains 26 figures,14 tables,and 80 references.