基于HMM的煤与瓦斯运动状态预测  

Prediction of coal and gas movement based on HMM state

作　　者：朱帅[1,2] 李长龙[1] 吴世跃[1] 

机构地区：[1]太原理工大学矿业工程学院,山西太原030024 [2]山西大同大学工学院,山西大同037003

出　　处：《西南民族大学学报（自然科学版）》2015年第4期507-511,共5页Journal of Southwest Minzu University(Natural Science Edition)

基　　金：国家科技支撑计划2007BAK29B01;山西省科技攻关项目(2007031120-02)

摘　　要：介绍了隐马尔科夫模型(HMM),并将该理论引入到煤与瓦斯运动状态预测的研究中,以工作面瓦斯浓度正常-值C、瓦斯浓度峰值P、峰值浓度与正常值之比B、两次采掘前后峰值浓度之比P的非接触式动态指标作为观察变量输入,煤与瓦斯的运动状态作为隐含变量输出,基于HMM提出一种新的预测煤与瓦斯运动状态的方法,通过实例训练预测,预测结果与实际监测趋势一致,且能实现提前预报.This article introduces hidden Markov model（HMM） and uses the theory to the study of coal and gas motion predic- tion. The following factors were input as observation variable：norms of gas concentration in working face, gas concentration peak, peak concentration and the ratio of the normal value, the ratio of the two peak concentration before and after the excavation of the non-contact dynamic index, and coal and gas movement state was output as implicit variable. A new method of predicting coal and gas motions is proposed based on HMM. Training instance results showed that the predicted results were consistent with actual monitoring trends, and could realize the forecast in advance.

关 键 词：隐马尔科夫模型 预报度 煤与瓦斯突出 危险性预测 

分 类 号：TD713[矿业工程—矿井通风与安全]