Research On Intelligent Evaluation Method Of Gas Extraction And Degassing In Goaf Based On Deep Learning

Gas drainage in goaf is the main means of mine gas control.The safe and efficient gas drainage effect evaluation plays an important role in the accurate extraction of mine gas.The intelligent evaluation of unloading gas drainage in goaf has important guiding significance for the unloading gas drainage project in goaf.This article is based on engineering data collection,theoretical analysis,model construction and training,prototype system design and development and field test,this paper proposes an evaluation index system for unloading gas drainage in goaf,and constructs long and short term memory network based on LSTM(Long Short Term Memory).The prediction model of the gas drainage and gas drainage evaluation index in the goaf of the network is formed,and the intelligent evaluation method of the gas drainage and drainage in the goaf is formed,and the pressure relief gas drainage evaluation system of the goaf is developed.The research work of the thesis are as follows:(1)On the basis of the principle and.comprehensive technical analysis of unloading gas drainage in goaf,analysis of key influencing factors such as drilling factors,wind gas concentration factors and extraction concentration factors,Considering the mutual coupling between various indicators and the impact on the comprehensive evaluation of gas drainage in goaf,the pressure extraction in the goaf such as gas drainage concentration,drainage flow and wind gas concentration is selected based on the analytic hierarchy process and relational matrix method.The evaluation index of gas drainage is used to establish the index satisfaction model by fuzzy comprehensive evaluation.The evaluation index system of gas drainage in the goaf is proposed,and the effect of gas drainage in the goaf is evaluated.(2)For the prediction accuracy of the gas drainage evaluation index in the goaf,the One-hot code is used to pre-process the extraction measurement data,reduce the data dimension,construct the data time window,and press the mine gas extraction measurement data.The 8:1:1 ratio is used to divide the data set,and a four-layer LSTM algorithm evaluation index system prediction model is constructed.By adjusting the parameters such as time step,loss function and optimization function,the accuracy and robustness of the model are improved.Compared with other predictive model algorithms,the LSTM model can solve the gradient disappearance problem with faster convergence speed and higher accuracy.(3)For the problem of intelligent classification of pressure-relief and gas extraction in goaf,the Lagrangian interpolation method and the average value correction method are used to pre-process the data of the extraction measurement data.The shallow machine learning evaluation method based on support vector machine and the deep learning evaluation method based on convolutional neural network are used to construct the intelligent evaluation model of unloading gas drainage in goaf.Compared with the support vector machine classification model of shallow neural network,the convolutional neural network classification model is more suitable for the intelligent evaluation of the unloading gas drainage in the goaf due to the superior learning ability of the deep neural network and the higher accuracy.(4)In order to realize the high-efficiency intelligent evaluation and visual display of the gas drainage in the goaf,the front-end interface and the back-end data structure of the intelligent evaluation system for the unloading gas drainage in the goaf are designed,and the algorithm integration and model are carried out on the cloud platform.The development and deployment of packaging and extraction evaluation and intelligent control system solves the key problems of data query and caching in the system development process,and provides software model and technical support for precision extraction of unloading gas in goaf.(5)According to the evaluation effect of gas drainage in the goaf,combined with the test data of high-level borehole gas drainage in the test mine,the effect of gas drainage in the goaf is analyzed,and the evaluation level of gas drainage is divided.According to the evaluation results,the recommended measures for the quality of drilling and sealing and the adjustment of suction negative pressure are proposed to realize the integrated process of pressure relief gas drainage monitoring,effect evaluation and intelligent regulation in the goaf,thus ensuring accurate and efficient extraction of pressure relief gas.Based on the above research results,field test verification is carried out,which forms an accurate and efficient intelligent evaluation method for unloading gas drainage in goaf,which provides a powerful basis for the evaluation of unloading gas drainage in goaf.