矿井涌水量动态无偏灰色马尔科夫预测  

Unbiased Grey Markov Prediction of Mine Inflow Dynamics

作　　者：王晶[1] 高国臣 WANG Jing;GAO Guo-cheng(Department of Resources and Mechanical Engineering,Lyuliang University,Lvliang 033000,China;Department of Computer Science and Technology,Lyuliang University,Lvliang 033000,China)

机构地区：[1]吕梁学院资源与机械工程系,山西吕梁033000 [2]吕梁学院计算机科学与技术系,山西吕梁033000

出　　处：《数学的实践与认识》2024年第9期233-240,共8页Mathematics in Practice and Theory

基　　金：吕梁市高新研发计划(2022GXYF10,2023GXYF16)。

摘　　要：矿井涌水量预测是防治煤矿水灾害发生的重要技术数据.为提高矿井涌水量预测精度,以传统灰色GM(1,1)模型为基础,构建无偏灰色马尔科夫模型,通过分析矿井2012-2017年矿井涌水量原始数据,预测2018-2019年矿井涌水量数据,利用无偏GM(1,1)模型代替传统的灰色GM(1,1)模型,通过拟合分析得出矿井涌水量数据变化特征,在此基础上采用马尔科夫模型对矿井涌水量进行预测,对原始数据进行更新,并对四种预测模型预测结果进行误差分析.研究结果表明:动态无偏灰色马尔科夫模型符合矿井涌水量数据特征,采用灰色理论和马尔科夫链处理动态数据具有明显的优势,消除传统模型自身固有误差,在预测值得基础上进一步优化,能够大大提高预测精度;动态无偏灰色马尔科夫模型预测误差比传统、无偏以及灰色马尔科夫预测模型的平均误差降低了0.072、0.064、0.0245,其矿井涌水量预测精度达到98.91%.Accurate prediction of mine water inflow is of great significance to mine safety production.To improve the predictive accuracy of mine water inflow in the traditional grey GM(1,1)model as the foundation,build the unbiased grey Markov model,through the analysis of the mine mine water inflow in 2012-2017 raw data,prediction of mine water inflow data from 2018 to 2019,using the unbiased GM(1,1)model instead of the traditional grey GM(1,1)model,through the fitting analysis of mine water inflow data variation characteristics,and on this basis,by using the Markov model to forecast the mine water inflow,updates to the original data,and the four kinds of prediction model to predict result error analysis.The results show that the dynamic unbiased gray Markov model accords with the characteristics of mine inflow data,and the use of gray theory and Markov chain to process dynamic data has obvious advantages,eliminate the inherent errors of the traditional model,and further optimize the model on the basis of the prediction value,which can greatly improve the prediction accuracy.The prediction error of dynamic unbiased gray Markov model is 0.072,0.064 and 0.0245 lower than the average error of traditional,unbiased and gray Markov model,and the prediction accuracy of mine inflow is 98.91%.

关 键 词：安全工程 矿井涌水量 动态无偏 灰色马尔科夫 

分 类 号：O211.6[理学—概率论与数理统计] TD742[理学—数学]