Research On The Soft Sensing Of Coal And Gas Outburst Based On Improved PCA-ELM

Coal and gas outburst is a serious dynamic disaster in mine, which seriously threatens the lives safety of personnel and safety in production in coal mine. Therefore, which is a key prevention and control objects of coal mine. Based on the full analysis of the mechanism and various factors of coal and gas outburst, the modeling method of soft sensor based on improved PCA-ELM algorithm is proposed in the paper. The new model was used to forecast coal and gas outburst.The following researches are expanded focusing on several key technical problems of soft sensor for coal and gas outburst forecast:(1) This paper fully analyses the mechanism and various factors of coal and gas outburst as well as discusses main factors which affect the accuracy of soft sensor for coal and gas outburst forecast. The results of the discussion are summarized as follows:data collection, data processing, auxiliary variables selection and modeling algorithm.(2) This paper put forward data preprocessing and auxiliary variables selection are two key technical problems, which should be careful study to improve the accuracy and feasibility of soft sensor for coal and gas outburst forecast. The collected data needs to be analysis and pretreatment. Principal component analysis theory is introduced to optimize the selection, reduce dimensionality to improve the accuracy and computational efficiency of the model.(3) Since the input weighs and the hidden layer thresholds of extreme learning machine are determined randomly, resulting in uncertainty on extreme learning machine. In this paper, particle swarm optimization algorithm is used to select the input weighs and the hidden layer thresholds of extreme learning machine, and the improved extreme learning machine algorithm is used for soft sensor modeling. The results showed that the soft sensor mode based on improved PCA-ELM algorithm had high measurement precision and generalization ability.