论文部分内容阅读
由于岩爆是深部地下工程常见的一种重大工程灾害,因此,岩爆预测研究具有重大的现实意义。岩爆影响因素众多且关系复杂,不能用简单的方法进行分析判断,一般在工程类比的基础上,采用聚类的方法进行。但由于岩爆问题环境的复杂性,岩爆预测的聚类问题是一个复杂的模糊、随机优化问题,采用传统方法难免带来很多局限性。为了更好地解决这类问题,首次把蚁群聚类算法这种新近提出的仿生聚类算法引入岩爆研究领域,以解决其预测问题,提出一种岩爆预测的新方法。该方法在分析岩爆实例资料的基础上,采用蚁群聚类算法,以工程类比的思想判断岩爆的发生状态。两个工程应用实例证明,该算法可以自动把岩爆事件分成几种类似的状态,判断准确率较高,计算速度较快,是一种比较实用的岩爆预测新方法,值得在岩石地下工程研究领域推广应用。
Because rockburst is a common major disaster in deep underground engineering, the research on rockburst prediction is of great practical significance. Rockburst has many influencing factors and complicated relations. It can not be analyzed and judged by a simple method. Generally, based on the engineering analogy, the method of clustering is used. However, due to the complexity of the rockburst environment, the clustering problem of rockburst prediction is a complex fuzzy and stochastic optimization problem, and the traditional methods inevitably bring about many limitations. In order to solve these problems better, for the first time, the newly proposed bionic clustering algorithm based on ant colony clustering algorithm is introduced into the field of rockburst research in order to solve its prediction problem and propose a new method of rockburst prediction. Based on the analysis of rock burst data, this method uses ant colony clustering algorithm to judge the occurrence of rock burst by engineering analogy. Two engineering examples prove that this algorithm can automatically separate the rockburst into several similar states, which has a higher accuracy and faster calculation speed. It is a practical new method for rockburst prediction and is worth to be used in rock underground engineering Research areas to promote the application.