Research On Calculation Model Of Mining Subsidence Prediction Parameters Based On Convolutional Neural Network

Surface movement caused by coal mining may cause surface subsidence,landslides,house collapse,damage to production and living facilities,etc.,and have a great impact on the safe production of mining areas and the lives and properties of surrounding residents.With the continuous advancement of the strategy of ecological civilization construction,the concept of "three under" mining and intelligent mine construction has made the improvement of the method of obtaining mining subsidence prediction parameters a problem that scholars urgently need to study.Regarding the complexity of the parameter influencing factors in the existing mining subsidence prediction methods,by analyzing the influence relationship between the mining subsidence prediction parameters and the geological mining factors,the main factors affecting the prediction parameters are determined.Principal Component Analysis(PCA)and Pearson's correlation coefficient method are used to determine the degree of correlation of the data,as far as possible to avoid the correlation between geological and mining factors and the error caused by the overlap of information on the prediction accuracy of mining subsidence parameters.And establish a regression model to confirm the correctness of the selected principal components.Established a convolutional neural network(Convolutional Neural Network,CNN)and PCAconvolutional neural network mining subsidence prediction parameter calculation model,the mining subsidence parameter predicted results and actual values are compared and analyzed.The results show that the data after the principal component analysis can produce fewer errors when applied to the convolutional neural network,which verifies the reliability and accuracy of the principal component data set used in the convolutional neural network.In order to further improve the accuracy of the model,the genetic algorithm(GA)is used to optimize the convolutional neural network,and the calculation model with strong self-organization,self-learning and adaptability is obtained in mechanism.The optimized convolutional neural network model is used to predict the principal component data,and the prediction results are compared and analyzed with the original data.The experimental results prove that the convolutional neural network with optimized algorithm has higher accuracy in parameter prediction.Through the case analysis of Bulianta Mine in Ordos and Sanjiaohe Mine in Shanxi,the results show that this model can predict mining subsidence under different mining conditions,provide reference for optimizing coal pillar design and land reclamation.It also has great theoretical significance and application value for disaster early warning evaluation and guidance for disaster prevention.The paper has 30 pictures,20 tables,and 67 references.