基于数字钻进技术和量子遗传-径向基函数神经网络的围岩类别超前识别技术研究  

Advanced prediction of surrounding rock classification based on digital drilling technology and QGA-RBF neural network

作　　者：邱道宏[1,2] 李术才[1] 薛翊国[1] 田昊[1] 闫茂旺 

机构地区：[1]山东大学岩土与结构工程研究中心,济南250061 [2]成都理工大学地质灾害防治与地质环境保护国家重点实验室,成都610059

出　　处：《岩土力学》2014年第7期2013-2018,共6页Rock and Soil Mechanics

基　　金：国家自然科学基金(No.51309144);地质灾害防治与地质环境保护国家重点实验室开放基金(No.SKLGP2013K019);山东大学自主创新基金(No.2012TS063)

摘　　要：围岩类别超前分类是隧道施工过程中必须开展的一项工作,其直接关系到后续的开挖及施工支护方案。为有效地进行隧道围岩类别超前分类,提出了基于数字钻进技术和量子遗传(QGA)-径向基函数(RBF)神经网络的围岩类别超前分类方法。以数字钻进技术为基础,从钻进参数中提取有用信息,构建围岩类别超前分类指标体系。采用量子计算原理对遗传算法进行改进,通过量子位编码和量子旋转门更新种群,以此来确定RBF神经网络的参数,建立了基于QGA-RBF神经网络的围岩类别超前识别系统。最后将该方法应用于青岛胶州湾海底隧道的围岩类别超前识别中,结果表明,该方法具有较高的准确性,其结果为围岩类别超前分类提供了一种新思路。Conducting the advanced surrounding rock classification is a necessary working in the process of tunnel construction, which is directly related to subsequent excavation and supporting construction scheme. In order to conduct surrounding rock classification ahead tunnel advancing effectively, the advanced surrounding rock classification method based on digital drilling technology and quantum genetic algorithm （QGA）-radical basis function （RBF） neural network is put forward. The method extracts useful information from the drilling parameters;and it establishes the indicators system of advanced surrounding rock classification. In the progress of establishing advanced surrounding rock classification index system based on QGA-RBF neural network, the genetic algorithm are improved by quantum calculation principle;and the parameters of RBF neural network could be determined by quantum bit and rotation gate renew population. Finally, the method is applied to the subsea tunnel across Qingdao Jiaozhou bay engineering. The results show that the method has higher prediction accuracy and provides a new idea in advanced prediction of surrounding rock classification.

关 键 词：围岩分类 超前识别 数字钻进 量子遗传算法(QGA) 径向基函数(RBF)神经网络 

分 类 号：O319.56[理学—一般力学与力学基础]