【作者】 梁栋；

【导师】 肖鹏； 袁晓翔；

【作者基本信息】 西安科技大学 ， 安全工程（专业学位）， 2018， 硕士

【摘要】 瓦斯事故可以称之为是关乎煤矿安全生产的首要难题,准确预测矿井瓦斯的涌出量,是有效地预防瓦斯灾害发生的基础。构建瓦斯涌出量预测模型,研究预测精度误差,能够有效实现煤矿瓦斯涌出量动态预测目标,对保障煤矿安全生产具有重要意义。论文以陕西澄合矿区某矿综采工作面为基础,根据5#煤层地质情况,分析瓦斯赋存影响因素的实际情况,基于灰色关联度分析方法,得到各个影响因素与瓦斯含量的关联度,得出影响该矿瓦斯含量大小的主要影响因素为煤层厚度和埋深。通过对比研究现有瓦斯涌出量预测方法现状,分析得出这些预测方法存在的不足之处,提出一种基于改进的蚁群算法和小波神经网络耦合(IACA-WNN)的瓦斯涌出量预测模型。基于2种主要影响因素,分别建立了煤层厚度和埋深综合作用下的神经网络(BP)、蚁群算法-BP神经网络(ACA-BP)、蚁群算法-小波神经网络(ACA-WNN)和IACA-WNN瓦斯涌出量预测模型,并给出了每种模型的算法编码和四种模型预测误差算法编码。通过MATLAB软件分别对4种预测模型算法进行计算,对比现场实际数据,得出IACA-WNN瓦斯涌出量预测模型的误差最小、精度最高、收敛效果最好。通过对该矿综采工作面瓦斯防治措施的分析,在现有措施的基础上结合前期的预测结果,提出了更加适合井下现场实际的综采工作面瓦斯防治技术措施和安全管理措施,通过实践证明,该措施对综采工作面的安全生产具有重要的指导意义。更多还原

【Abstract】 Gas accident is the first problem that affects the safety production of coal mine.Accurate prediction of gas emission in mine is the foundation of effective prevention and control of gas disaster.It is of great significance to ensure the safety production of coal mine by constructing the prediction model of gas emission quantity and can effectively realize the dynamic prediction target of gas emission in coal mine.On the basis of a fully mechanized mining face in Shaanxi Chenghe,according to the 5＃coal bed geological situation,analysis of the actual situation of gas occurrence factors,based on grey relational analysis method,obtained the correlation degree between each influence factor and gas content,indicates that the main factors affecting the mine gas amount for thickness and buried depth of coal seam.By comparing the present research situation of the existing gas emission prediction method and analysising the deficiency of these prediction methods,puts forward a kind of based on improved ant colony algorithm and wavelet neural network coupling(IACA-WNN)of gas emission prediction model.Based on two main control factors,respectively established integrated neural network(BP)and ant colony algorithm-the BP neural network(ACA-BP),ant colony algorithm-wavelet neural network(ACA-WNN)and IACA-WNN gas emission prediction model under the action of the coal seam thickness and buried depth,and given the algorithm coding of each model and four models prediction error algorithm coding.By using MATLAB software to calculate the four prediction model algorithms respectively,and compared the actual data in the spot,it is concluded that the error of IACA-WNN gas emission prediction model is the smallest,the highest precision and the best convergence effect.Through the analysis of the mine gas prevention and control measures of fully mechanized working face,On the basis of existing measures and combined with previous prediction results,put forward more suitable for actual underground gas prevention and control technology measures and safety management measures of fully mechanized working face,through the practice has proved that the measures has an important guiding significance to the safety production of fully mechanized working face.更多还原