新信息优先原则下矿井回采工作面瓦斯涌出量的MUBGM(1,1)-Markov预测

1.西安科技大学 安全科学与工程学院,陕西 西安 710054; 2.西安科技大学 材料科学与工程学院,陕西 西安 710054; 3.中国科学院 信息工程研究所,北京 100093; 4.教育部 西部矿井开采及灾害防治重点实验室,陕西 西安 710054

GM(1,1)模型; 无偏; 马尔可夫; 瓦斯涌出量; 等维新息

Prediction of the working face gas emission quantity based on MUBGM(1,1)-Markov under new information priority principle
TIAN Shui-cheng1,4,YANG Xue-jian1,4,ZHAO Na-ying2,WANG Min3

(1.College of Safety Science and Engineering,Xi'an University of Science and Technology,Xi'an 710054,China; 2.College of Materials Science and Engineering,Xi'an University of Science and Technology,Xi'an 710054,China; 3.Institute of Information Engineering,Chinese Academy of Sciences,Beijing 100093,China; 4.Key Laboratory of Western Mine Exploration and Hazard Prevention,Ministry of Education,Xi'an 710054,China)

GM(1,1)model; unbiased; Markov; gas emission quantity; metabolic

DOI: 10.13800/j.cnki.xakjdxxb.2017.0608

备注

为实现煤矿瓦斯涌出量的高精度预测,进而对矿井通风设计和瓦斯事故防治提供理论依据,以传统GM(1,1)模型为基础,建立了一种等维新息式无偏灰色马尔可夫预测模型(MUBGM(1,1)-Markov模型)。该模型将无偏灰色GM(1,1)模型与马尔可夫模型相结合,依据2001—2010年西北地区某煤矿二矿区矿井相对瓦斯涌出量数据,在新信息优先原则下经等维新息处理和原始数据序列的不断更新,预测出2011—2016年该矿的相对瓦斯涌出量。研究结果表明,MUBGM(1,1)-Markov预测模型不仅消除了传统灰色GM(1,1)模型的固有偏差,而且预测值可随数据的变化实时更新,预测效果理想,2013—2015年的平均相对误差仅为0.012 4,在中长期预测中具备明显优势,其较高的准确性和较强的适用性,为矿井瓦斯涌出量的高精度预测提供了可靠的保障。

To achieve high precision prediction of coal mine gas emission quantity,and provide theoretical basis for mine ventilation design and gas accident prevention and control,we established a metabolic unbiased gray markov model(MUBGM(1,1)-Markov model)based on the traditional GM(1,1)model.According to the relative gas emission data of the No.2 mining area of a coal mine in the northwest region in 2001—2010,the model combined unbiased grey GM(1,1)model with Markov model to forecast the relative gas emission of the mine in 2011-2016 through the process of metabolism and continuous update of the original data sequence under new information priority principle.Research results show that MUBGM(1,1)-Markov model not only eliminates the inherent deviation of the traditional gray GM(1,1)model,but also updates the predictive value in real time with the change of the data,and the prediction effect is ideal.The average relative error of the model in 2013—2015 is only 0.012 4,which has obvious advantages in medium and long term prediction,and its higher accuracy and strong applicability provide a reliable guarantee for the high precision prediction of mine gas emission quantity.