摘要
综采工作面上隅角瓦斯浓度超限严重影响煤矿的安全生产,提高上隅角瓦斯浓度预测精度对瓦斯超限预警具有重要作用。以某煤矿某综采工作面上隅角9000个瓦斯浓度数据为研究对象,首先利用Matlab对数据进行缺失值、异常值、归一化处理,然后选择长短期记忆神经网络(LSTM)、卷积长短期记忆神经网络(CNN-LSTM)、卷积双向长短期记忆神经网络(CNN-BiLSTM)对处理后数据进行训练与学习,构建不同上隅角瓦斯浓度预测模型,最后采用均方根误差(RMSE)和平均绝对误差(MAE)作为预测模型的评价指标。结果表明:预测样本数据为9000时,LSTM模型、CNN-LSTM模型、CNN-BiLSTM模型的MAE分别为0.07607、0.065432、0.050242,RMSE分别为0.11959、0.10308、0.094238,对比LSTM模型与CNN-LSTM模型,CNN-BiLSTM模型预测平均绝对误差提升23.3%和34%,均方根误差提升8.6%和22.2%,因此CNN-BiLSTM模型预测精度更高,具有更好的可靠性。
The overrun of the upper corner gas concentration on the fully mechanized mining face seriously affects the safety production of coal mines,and improving the prediction accuracy of the upper corner gas concentration plays an important role in the early warning of gas overrun.Taking the data of 9000 gas concentrations in the upper corner of a fully mechanized mining face of a coal mine as the research object,firstly,Matlab was used to process the missing values,outliers and normalization of the data,and then the long short-term memory neural network(LSTM),convolutional long short-term memory neural network(CNN-LSTM)and convolutional bidirectional long short-term memory neural network(CNN-BiLSTM)were selected to train and learn the processed data,and the prediction model of gas concentration in different upper corners was constructed.Finally,the root mean square error(RMSE)and the mean absolute error(MAE)were adopted as the evaluation indicators of the prediction model.The results show that when the predicted sample data is 9000,the MAE of the LSTM model,the CNN-LSTM model and the CNN-BiLSTM model are 0.07607,0.065432 and 0.050242,respectively,and the RMSE is 0.11959,0.10308 and 0.094238,respectively,and the average absolute error of the CNN-BiLSTM model is increased by 23.3%and 34%compared with the CNN-LSTM model.The root mean square error is increased by 8.6%and 22.2%,so the CNN-BiLSTM model has higher prediction accuracy and better reliability.
作者
胡双虎
HU Shuanghu(Shanxi Shouyang Luyang Xiangsheng Coal Industry Co.,Ltd.,Jinzhong 045400,China)
出处
《煤》
2025年第8期39-43,共5页
Coal
作者简介
胡双虎(1972-),男,山西襄垣人,工程师,从事通风安全、防灭火、地质防治水工作,E-mail:2581531015@qq.com。
