基于人工智能技术的矿井水害来源识别模型库的建立方法研究  

Research on Establishment Method of Mine Water Source Identification Model Base Based on Artificial Intelligence

作　　者：阴宁宝 郝军 余生晨 YIN Nin-bao HAO Jun YU Sheng-chen(Shanxi Lu＇an Environmental Energy Development Co.Ltd , Changzhi, 046000, China Yu＇wu Coal Industry Co.Ltd., Changzhi 221000, China School of Computer, North China Institute of Science and Technology, Yanjiao, 065201, China)

机构地区：[1]山西潞安环保能源开发股份有限公司,山西长治046000 [2]山西潞安环保能源开发股份有限公司余吾煤业有限公司,山西长治221000 [3]华北科技学院计算机学院,北京东燕郊065201

出　　处：《华北科技学院学报》2017年第4期24-28,共5页Journal of North China Institute of Science and Technology

基　　金：中央高校基本科研业务费资助(JSJ1207B;3142013093)

摘　　要：为了建立合理的矿井水害识别模型库,以便提高水害来源识别的准确率,提出了建立水害识别模型库(水的水化学模型)的原则和方法技术。这个原则是各类水害来源模型内部的类内离散度尽可能小,使其具有代表性;各个模型之间的类间离散度尽可能大,以便清楚的区分各个水害来源的总原则。给出了描述这个原则的数学公式以及实现上述原则的方法技术。在山西潞安环保能源开发股份有限公司的多个煤矿进行了生产性验证。生产实践证明提出的原则和方法技术是可行的、识别矿井水的来源(判定水的类型)的准确率可达到95%。In order to establish a reasonable model library of mine water disaster,and to improve the accuracy of water source identification,the principles,methods and techniques of establishing model library of water disaster identification are presented.The principles is that the intra class dispersion of each water source model is as small as possible,which makes it representative and the inter class dispersion between the models is as large as possible in order to clearly distinguish each source of water damage.The mathematical formulas for describing this principle and the methods and techniques for realizing the above principles are given.Production verification has been carried out in the coal mine owned by Shanxi Lu＇an environmental energy development Co.Ltd.Experimental results and production practice show that the principle and method is efficient and feasible,and the detection right rate of flood waters was above 95% and the method is efficient and feasible.

关 键 词：矿山水化学模型库 水源识别 类内离散度 类间离散度 

分 类 号：TD745.21[矿业工程—矿井通风与安全]