矿井相对瓦斯涌出量动态无偏灰色马尔科夫预测  

Application of the Markov bias-free dynamic grey model for the relative gas gush-out rate predict

作　　者：熊祖强[1,2] 王晓蕾[3] 

机构地区：[1]河南理工大学能源科学与工程学院,河南焦作454003 [2]煤炭安全生产河南省协同创新中心,河南焦作454003 [3]西南石油大学地球科学与技术学院,成都610500

出　　处：《安全与环境学报》2015年第3期15-18,共4页Journal of Safety and Environment

基　　金：国家自然科学基金面上基金项目(51274088);国家自然科学基金青年科学基金项目(50804013)

摘　　要：矿井瓦斯涌出量预测对于煤矿瓦斯防治具有重要意义。为预测矿井瓦斯相对涌出量,以传统灰色GM(1,1)模型为基础,构建动态无偏灰色马尔科夫模型,通过分析潞安矿区某矿2003—2010年的煤矿相对瓦斯涌出量数据,预测2011—2012年煤矿相对瓦斯涌出量数据,利用无偏灰色GM(1,1)模型代替传统灰色GM(1,1)模型,通过拟合得到煤矿相对瓦斯涌出量数据变化趋势,并在此基础上利用马尔科夫模型进行预测,并在此预测中进行原始数据更新,并对4种预测方法的预测结果进行对比分析。结果表明,动态无偏灰色马尔科夫模型不但能够消除传统灰色GM(1,1)模型自身的固有偏差,而且能提高预测精度,平均绝对误差为3.2%,平均相对误差为2.59%,均低于传统灰色GM(1,1)模型与一般灰色马尔科夫模型。动态无偏灰色马尔科夫模型对于煤矿相对瓦斯涌出量数据的平均预测精度达到96.74%。The present paper intends to introduce its study results on how to apply Markov bias-free dynamic grey model for predicting the relative gas gush-out rate,which is of great significance for the gas-gush-out prevention and control in coal mining. For the research purpose,we have established a dynamic bias-free grey model DUGM-Markov on the basis of the traditional grey GM（ 1,1） one,which has been used for predicting the relative gas gush-out rate from 2011 to 2012 based on the research findings in the field from the period between 2003 to 2010. Instead of the traditional grey GM（ 1,1） model,the improved bias-free grey GM（ 1,1） model we have improved on has been trial-used and proved fit for the changing tendency of the relative gas gush-out rate in the coal mine production in predicting the fluctuation.Furthermore,we have also made the original data updated and made an analysis of the results of the four predicting methods through comparison. Thus,the dynamic bias-free grey Markov model has been made capable for getting rid of the inherent bias of the traditional GM（ 1,1） model with a much higher predicting accuracy,particularly for the cases of middle and long-term predict than it can be achieved by using the traditional grey GM（ 1,1） model. And,more specifically speaking,it has made it possible for the innovated model to reduce the average absolute error from 3. 2% to 2. 59% via the improved method we have proposed. And,alternatively speaking,we have successfully promoted the average prediction accuracy of the dynamic bias-free grey Markov model on the relative amount of the mining gas gush-out by over 96. 74%. In addition,as compared with the model of traditional grey Markov,the improved dynamic bias-free grey Markov model has been constructed on the latest basis and in a more harmonious accord with the actual data changes. Thus,it can be said that the method has been made more convenient and efficient in predicting the amount of coal gas gush rate by saving large amounts of complicated deri

关 键 词：安全工程 马尔科夫 动态 无偏 相对瓦斯涌出量 

分 类 号：X914[环境科学与工程—安全科学]