综掘面风筒出风口风流变化下的瓦斯预测神经网络模型  

Gas Concentration Prediction Model Based on Neural Network Under Airflow Changes of Air Duct Outlet in Fully Mechanized Heading Face

作　　者：龚晓燕[1] 孙康 侯翼杰 张欣怡[1] 李莹 夏治新 GONG Xiaoyan;SUN Kang;HOU Yijie;ZHANG Xinyi;LI Ying;XIA Zhixin(School of Mechanical Engineering,Xi'an University of Science and Technology, Xi'an,Shanxi 710054,China)

机构地区：[1]西安科技大学机械工程学院,陕西西安710054

出　　处：《矿业研究与开发》2019年第1期129-133,共5页Mining Research and Development

基　　金：国家自然科学基金面上资助项目(51874235);陕西省重点研发计划资助项目(2017GY-170)

摘　　要：综掘面长距离大断面掘进过程中,风筒出风口风流不能动态变化,导致瓦斯运移分布不合理,瓦斯爆炸风险大。为此,对出风口参数变化下的瓦斯浓度分布及预测模型进行研究。以柠条塔综掘面为对象,通过分析,确定了神经网络结构模型的输入层及隐含层数目,建立了出风口参数变化下的掘进端头瓦斯浓度预测模型,通过均方误差和拟合分析验证了神经网络预测模型精度,得出了风筒出风口距掘进端头5m和10m时,最佳排瓦效果的出风口调控方案,调控后瓦斯浓度分别降低了10%和11.6%,为煤矿瓦斯灾害的准确预测与及时进行瓦斯调控提供了依据。The airflow at the air duct outlet cannot be changed dynamically during the long-distance and large-section tunneling at the fully mechanized heading face,resulting in unreasonable distribution of gas distribution and large explosion risk.Therefore,the gas concentration distribution and prediction model under the variation of air outlet parameters were studied.Taking the fully mechanized heading face in Ningtiaota as an object,the input layer and the number of hidden layer for the structure model based on neural network were determined by analysis,and the prediction model of gas concentration at fully mechanized heading face under the changing outlet parameters was established.Then,the prediction accuracy of the neural network prediction model were verified by using the mean square error and fitting analysis.The optimal control scheme of the air outlet was obtained when the air outlet was 5 mand 10 maway from the heading face.After the implement of control scheme,the gas concentration was reduced by 10% and 11.6%respectively.The study provided a basis for accurate prediction and timely gas control of coal mine gas disasters.

关 键 词：综掘面 BP神经网络 风筒出风口 预测模型 调控方案 

分 类 号：TD263.5[矿业工程—矿井建设]