A Study a CNN-LSTM Model for coal gas concentration prediction model based on wavelet analysis
TU Xinning
YU Xingchen
LIU Dequan
WANG Yang
TONG Bo
Abstract：The high-precision gas concentration prediction model can accurately predict the change of gas con-centration in the mine and effectively prevent safety accidents caused by gas accumulation.Due to the low ac-curacy of the existing prediction model,it is difficult to deal with noise interference.Therefore,the actual working face of a coal mine in China was selected as the experimental scene,and the CNN-LSTM neural net-work model based on wavelet analysis was constructed to reduce noise interference and to realize accurate pre-diction of gas concentration.The same dataset was employed for training the LSTM model,the CNN-LSTM model and the wavelet analysis-based CNN-LSTM model respectively.The efficacy of each model was evalua-ted through the introduction of a loss function.The experimental results show that the Root Mean Square Error(RMSE)of CNN-LSTM gas concentration prediction model based on wavelet analysis is 3.02％,0.63％,0.6％,which are lower than those of LSTM model without wavelet analysis,CNN-LSTM model without wavelet analysis,and LSTM model based on wavelet analysis,respectively.The Mean Absolute Error(MAE)is re-duced by 2.62％,0.54％,0.56％respectively.It can be seen that compared with traditional prediction mod-els,the CNN-LSTM gas concentration prediction model based on wavelet analysis can effectively handle com-plex fluctuations and noise interference in gas time series data,exhibiting higher prediction accuracy and ro-bustness,thus providing reliable technical support for preventing gas disasters.
Keywords：coal mine safetygas concentration predictionwavelet analysisconvolutional neural networks(CNN)long short-term memory(LSTM)
Publication Date：2025-06-30
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：10( 21-30 )