摘要
煤矿瓦斯浓度精准预测及提早预警对于预防瓦斯灾害发生至关重要。为充分利用井下多传感器监测信息提升矿井瓦斯浓度预测及预警模型的性能,提出一种基于主成分分析(PCA)、门控循环单元(GRU)神经网络和支持向量机(SVM)组合的多参量瓦斯浓度预测及预警模型。针对监测数据的时序性、动态性和非线性强等特性问题,采用邻近均值法、小波降噪和归一化法对数据进行处理,利用PCA对数据降维以简化GRU模型拓扑结构,提高瓦斯浓度预测精度,通过构建基于SVM的矿井瓦斯浓度预警模型实现对矿井安全状态的实时动态监测。选取安徽某煤矿171105工作面的实测数据对PCA-GRU-SVM模型预测结果与性能进行验证。仿真结果表明:相对于PCA-LSTM、PCA-RF和PCA-BP模型,本文构建的预测模型的平均绝对误差(MAE)分别减少了18.45%、56.36%和87.3%,均方根误差(RMSE)分别减少了5.17%、9.04%和67.52%,预警模型的预测准确率为94.1%,说明该模型具有较高的拟合度和预测精度。该研究结果可为实现瓦斯灾害的预测及超前预警提供参考,对矿业安全生产具有重要意义。
Accurate prediction and early warning of coal mine gas concentration are very important for preventing gas disasters.In order to make full use of underground multi-sensor monitoring information to improve the performance of gas concentration prediction and early warning model,a gas prediction and early warning model based on the combination of principal component analysis(PCA),gate recurrent unit(GRU)neural network and support vector machine(SVM)was proposed.Aiming at the problems of time series,dynamics and strong nonlinearity of data,firstly,the adjacent mean method,wavelet noise reduction and normalization method were used to process the data,and PCA was used to reduce the dimension of the data to simplify the topological structure of GRU network model and improve the prediction accuracy.Finally,a mine gas concentration early warning model based on SVM was constructed to realize real-time dynamic monitoring of mine safety status.The PCA-GRU-SVM model was verified by the measured data of 171105 working face in a mine in Anhui province.The simulation results show that compared with PCA-LSTM,PCA-RF and PCA-BP models,the average absolute error(MAE)of this model is reduced by 18.45%,56.36%and 87.3%respectively,the root mean square error(RMSE)is reduced by 5.17%,9.04%and 67.52%respectively,and the prediction accuracy of the early warning model is 94.1%.The model method proposed in this study provides a reference for the prediction and early warning of gas disasters,and is of great significance to the safe production of mining industry.
作者
秦岩
盛武
QIN Yan;SHENG Wu(School of Economics and Management,Anhui University of Science&Technology,Huainan 232001,China)
出处
《安全与环境工程》
CAS
CSCD
北大核心
2023年第6期81-88,共8页
Safety and Environmental Engineering
基金
安徽省高校研究生科研项目(YJS20210411)
国家自然科学基金项目(71971003)
安徽省自然科学基金项目(1808085MG212)。
作者简介
秦岩(1995),女,硕士研究生,主要研究方向为矿业安全与数据分析。E-mail:965653946@qq.com;通讯作者:盛武(1969),男,博士,副教授,主要从事矿业安全与数据挖掘等方面的研究。E-mail:604597010@qq.com。
