基于神经网络的节理岩体单轴强度预测  

Uniaxial Strength Prediction of Jointed Rock Based on Neural Network

作　　者：胡安龙[1] 王孔伟[1] 邓华锋[1] 常德龙[1] 肖志勇[1] 李春波[1] 

机构地区：[1]三峡大学土木与建筑学院,443002

出　　处：《灾害与防治工程》2015年第1期42-48,共7页Disaster and Control Engineering

基　　金：国家自然科学基金资助项目（51309141）;国家自然科学基金资助项目（51279091）;水利部公益基金项目（201401029

摘　　要：基于RMT-150C岩石力学试验系统上的节理岩体单轴压缩试验结果，本文分析了影响节理岩体单轴压缩强度因素。节理岩体的单轴压缩强度与节理贯通度、节理倾角、节理个数等因素有关，且它们是一种复杂的非线性关系。同一贯通度节理岩体在节理倾角为0。时峰值强度最大；同一节理倾角岩体的峰值强度随着贯通度的增加而减小。考虑到多种因素对节理岩体单轴压缩强度的影响，本文建立了BP神经网络模型以对节理岩体的单轴压缩强度进行预测，然后利用遗传算法优化BP神经网络模型。通过对岩石单轴压缩试验样本数据的学习，遗传算法优化的神经网络模型能够很好预测节理岩体的单轴压缩强度。该方法可以推广到岩石本构模型的研究中。Based on results of the jointed rock mass single-axis compression test with RMT-150C rock mechanics test system, this paper analyzes the factors affecting the jointed rock uniaxial compressive strength. The uniaxial compressive strength of rock joints has complex nonlinear relationship with the joint through degrees, inclination of joints, and number of joints. The uniaxial compressive strength of the same joints through degree has maximum intensity when the inclination is 0°, and the intensity decreases with the increase of the joint through degree. Taking into account a variety of factors that affect the uniaxial compressive strength of jointed rock mass, this paper establishes a BP neural network model for jointed rock uniaxial compressive strength prediction and then takes advantage of genetic algorithm to optimize BP neural network model. Compression test data through u-niaxial learning, genetic algorithm optimization neural network model can uniaxial compressive strength of jointed rock masses. This method can be study of rock constitutive model.

关 键 词：节理岩体 遗传算法 BP神经网络 单轴压缩强度 

分 类 号：TU452[建筑科学—岩土工程]