采矿与安全工程学报 ›› 2014, Vol. 31 ›› Issue (2): 196-202.

• 论文 • 上一篇    下一篇

基于聚类分析的微震定位二次优化研究

  

  1. 1.北京科技大学土木与环境工程学院,北京 100083;2.北京安科兴业科技有限公司,北京 100083
  • 出版日期:2014-03-15 发布日期:2014-03-20
  • 作者简介:朱权洁(1984-),男,湖北省武汉市人,博士研究生,从事微震监测、矿山压力及岩层控制方面的研究。 E-mail:youyicun2008@gmail.com Tel:13718459909
  • 基金资助:

    国家重点基础研究发展计划(973)项目(2010CB226803);国家自然科学基金项目(51174016,51274022,51204010)

Quadratic optimization of mining microseismic source positioning based on cluster analysis

  • Online:2014-03-15 Published:2014-03-20

摘要: 为了提高微震定位的精度,利用MATLAB软件及其工具箱编制相关分析、处理程序,进行了关于微震定位结果的二次优化研究。利用传统数学方法,将微震定位问题简化为求解线性方程组,建立基于多传感器的四四组合定位模型,四四一组求解其定位结果;并采用k-means聚类方法对组合定位结果进行优化处理,减小定位结果的奇异性。设定核心簇和外围簇2类,引入聚心曲线拐点及噪声偏移距离概念,建立聚类继续与否的判定准则。通过某矿现场验证表明,优化后的结果与真实震源误差为(1.8,2.83,11.6),空间距离约为12.08 m。该方法算法简单,操作方便,且精度较之单一定位方法有所提高,基本满足现场需求。

关键词: 微震定位, 四四组合定位, 二次优化, 聚类分析

Abstract: In order to improve the precision of the microseismic location, the quadratic optimization of mining microseismic source location has been researched through using MATLAB software and its toolbox in this paper. First, the microseismic location problem is boiled down to solving linear equations using traditional mathematics method and a Four-Four combined positioning model is set based on a multiple sensors; second, the k-means clustering method is applied to optimize the positioning results. Two classes—core cluster and peripheral cluster are set to processing those data. When it meets the principle of heart knee point and noise migration distance, the circular will stop and the final result will be solved. The result of field test shows that the optimized results are in close proximity to the real source; the distance between those is about 12.08 m (x:1.8 m,y:2.83 m,z:11.6 m). This method is simple and easy to operate, and the precision has been improved compared with the single positioning method.

Key words: microseismic location, Four-Four method, quadratic optimization, cluster analysis