摘要
煤层预抽后残余瓦斯含量过高致使工作面回采过程中瓦斯突出及瓦斯涌出现象频发,存在极大瓦斯爆炸、火灾等隐患.为有效控制煤层预抽后工作面残余瓦斯含量,以林华煤矿为例建立9#煤层预抽后工作面残余瓦斯含量预测模型.在粗糙集等理论基础上利用Weka软件,通过SVM分类机、BP神经网络和NaiveB ayes分类器,分析验证各因素对残余瓦斯含量的影响,获得12组待测样本预测结果的准确率;随之进行详细精度,混淆矩阵和节点错误率分析,得出SVM分类机在瓦斯残余含量的预测上最为精准;最后通过MATLAB软件实现PSO算法对SVM的优化,提高预测精准度,得出最终预测结果.
Excessive gas content after pre-pumping of coal seams leads to frequent gas outbursts and gas surges in the working face mining process,and there are hidden dangers such as gas explosion and fire.In order to effectively control the residual gas content of the coal face after pre-extraction of coal seam,the prediction model of residual gas content in the working face of 9#coal seam pre-extraction was established by taking Linhua Coal Mine as an example.Based on the theory of rough set and so on,we use Weka software to analyze and verify the influence of various factors on residual gas content through SVM classifier,BP neural network and NaiveBayes classifier.The accuracy of the prediction results of 12 groups of samples to be tested is obtained;followed by detailed accuracy,confusion matrix and node error rate analysis,it is concluded that the SVM classifier is the most accurate in predicting gas residual content;Finally,the MATLAB software is used to optimize the SVM by the PSO algorithm,and the prediction accuracy is improved,and the final prediction result is obtained.
作者
王新
郝建
陈军
程为佳
WANG Xin;HAO Jian;CHEN Jun;CHENG Wei-jia(State Key Laboratory Breeding Base for Mining Disaster Prevention and Control,Shandong University of Science and Technology,Qingdao 266590,China;College of Mining and Safety Engineering,Shandong University of Science and Technology,Qingdao 266590,China)
出处
《数学的实践与认识》
北大核心
2020年第9期97-105,共9页
Mathematics in Practice and Theory
基金
国家自然科学基金(51804180,51574055)
山东省重点研发计划(GG201710010020)。
