东北大学学报(自然科学版) ›› 2024, Vol. 45 ›› Issue (2): 244-251.DOI: 10.12068/j.issn.1005-3026.2024.02.012

• 资源与土木工程 • 上一篇    

基于D-S证据理论的岩爆预测方法研究

高永涛1, 朱强1, 吴顺川1,2, 王勇兵3   

  1. 1.北京科技大学 土木与资源工程学院,北京 100083
    2.昆明理工大学 国土资源工程学院,云南 昆明 650093
    3.云南驰宏锌锗股份有限公司,云南 曲靖 655000
  • 收稿日期:2023-03-03 出版日期:2024-02-15 发布日期:2024-05-14
  • 作者简介:高永涛(1962-),男,山东威海人,北京科技大学教授,博士生导师.
  • 基金资助:
    国家自然科学基金资助项目(51934003)

A Study of Rockburst Prediction Method Based on D-S Evidence Theory

Yong-tao GAO1, Qiang ZHU1, Shun-chuan WU1,2, Yong-bing WANG3   

  1. 1.School of Civil and Resource Engineering,University of Science and Technology Beijing,Beijing 100083,China
    2.Faculty of Land Resources Engineering,Kunming University of Science and Technology,Kunming 650093,China
    3.Yunnan Chihong Zinc and Germanium Co. ,Ltd. ,Qujing 655000,China. Corresponding author: ZHU Qiang,E-mail: qiangzhus@163. com
  • Received:2023-03-03 Online:2024-02-15 Published:2024-05-14

摘要:

为了有效预测岩爆,提出基于D-S证据理论的岩爆预测方法.首先,选取与岩爆发生相关的6个指标因素作为证据体,并通过模糊物元框架和正态型隶属度函数构建证据体的基本概率分配.然后,利用K均值将证据体分类,并提出簇内证据用传统方式融合而簇间证据用权重方式融合的组合融合规则,以减轻高冲突证据融合的不利影响.最后,将模型应用在秦岭终南山公路隧道2号竖井工程,且与经验方法对比.为了分析预测过程的不确定性和估计岩爆发生概率,采用蒙特卡洛模拟进行抽样仿真,并通过Spearman秩相关系数衡量输入指标的全局敏感性.研究结果表明:输入指标在不同的岩爆案例的影响程度差异较大且方向不同;5个岩爆案例的发生概率在40.8%~70.1%之间.该模型表现出优异的预测分类性能,可为深埋地下工程岩爆预测提供参考.

关键词: 岩石力学, 岩爆预测, D-S证据理论, 模糊物元, K均值

Abstract:

To effectively predict rockburst, a rockburst prediction method based on D-S evidence theory is proposed. Firstly, six indicator factors related to rockburst occurrence are selected as evidence, and the basic probability assignment of evidence is constructed through fuzzy matter-element framework and normal type degree of membership function. Then, the evidence is classified using K-means and a combined fusion rule of fusing evidence within clusters in classical Dempster’s rule while fusing evidence between clusters in the weighted way is proposed to reduce the adverse effects of high conflict evidence fusion. Finally, the model is applied to the No.2 shaft project of the Qinling Zhongnanshan highway tunnel and compared with the empirical method. To address the uncertainty in the prediction process and to estimate the probability of rockburst occurrence, Monte Carlo simulation is used to perform sampling simulations and the global sensitivity of the input indicators is measured by Spearman’s rank correlation coefficient. The results show that the influence of the input indicators varies widely and in different directions for different rockburst cases; the probability of occurrence of the five rockburst cases ranges from 40.8% to 70.1%. The model shows outstanding predictive classification performance and can provide reference for rockburst prediction in deeply buried underground projects.

Key words: rock mechanics, rockburst prediction, D-S evidence theory, fuzzy matter-element, K-means

中图分类号: