石油学报 ›› 2014, Vol. 35 ›› Issue (6): 1172-1181.DOI: 10.7623/syxb201406015

• 油田开发 • 上一篇    下一篇

水力压裂微地震粒子群差分进化定位算法

盛冠群1, 李振春1, 王维波1, 王振宇1, 朱海伟2, 彭国民1   

  1. 1. 中国石油大学地球科学与技术学院 山东青岛 266555;
    2. 中国石油化工股份有限公司石油物探技术研究院 江苏南京 210000
  • 收稿日期:2014-02-27 修回日期:2014-08-19 出版日期:2014-11-25 发布日期:2014-10-13
  • 通讯作者: 盛冠群,男,1987年8月生,2010年获长江大学学士学位,现为中国石油大学(华东)地球科学与技术学院博士研究生,主要从事微地震信号处理与定位方法的研究。Email:88407581@qq.com
  • 作者简介:盛冠群,男,1987年8月生,2010年获长江大学学士学位,现为中国石油大学(华东)地球科学与技术学院博士研究生,主要从事微地震信号处理与定位方法的研究。Email:88407581@qq.com
  • 基金资助:

    国家高技术研究发展计划(863)项目(2011AA060302,2011AA060301)和中央高校基本科研业务费专项资金项目(13CX02098A)资助。

A source location method for microseismic monitoring based on particle swarm optimization combined with differential evolution algorithm

Sheng Guanqun1, Li Zhenchun1, Wang Weibo1, Wang Zhenyu1, Zhu Haiwei2, Peng Guomin1   

  1. 1. School of Geosciences, China University of Petroleum, Shandong Qingdao 266555, China;
    2. Sinopec Petroleum Geophysical Technology Research Institute, Jiangsu Nanjing, 210000, China
  • Received:2014-02-27 Revised:2014-08-19 Online:2014-11-25 Published:2014-10-13

摘要:

为了提高微地震在初至时间不准确条件下的定位精度,通过研究理论模型下对初至施加不同程度的扰动时震源定位的影响,讨论了由于粒子群定位方法的发散性而导致的定位效果偏差问题。把粒子群算法全局搜索速度快的特点与差分进化方法相结合,将种群中任意两个个体差分变异后,进行杂交、贪婪选择操作,生成新的个体,从而增加了种群的多样性,改善了震源定位的发散性,提高了时差定位的精度,最终形成了微地震粒子群差分进化定位算法。通过理论模型测试,表明微地震粒子群差分进化定位算法针对有扰动的初至时间具有更高精度,且震源定位的发散性较传统粒子群定位得到了改善。实际微地震资料反演结果也进一步验证了微地震粒子群差分进化定位算法的应用效果。

关键词: 压裂, 微地震, 粒子群算法, 差分进化算法, 震源定位

Abstract:

This study aims to improve the accuracy of source location in case of inaccurate first arrival time. It investigates the influence of disturbances to the theoretical first arrival on source location, and then discuses the deviation of source location due to the divergence of particle swarm optimization (PSO) algorithm. The PSO algorithm characterized by high-speed global search is combined with differential evolution algorithm. After differential mutation of any two individuals in the population, new individuals are generated by hybridization and greedy selection operation. This process increases the population diversity, reduces the divergence of source location, and improves the accuracy of source location. Ultimately, a source location method based on the PSO and differential evolution algorithms is established for microseismic monitoring. Test data from the theoretical model show that the proposed method has higher accuracy in case of a disturbed first arrival time, whereas the divergence of source location is reduced compared with traditional PSO method. The applied effect of the established method is further verified by inversion of the actual microseismic data.

Key words: fracturing, microseismic, particle swarm optimization algorithm, differential evolution algorithm, source location

中图分类号: