水利水电技术2017，Vol.48Issue(1)：133-138,6.DOI:10.13928/j.cnki.wrahe.2017.01.025
基于改进向量机的岩体质量分级研究
Improved support vector machine-based study on rock mass quality classification
摘要
Abstract
Based on the engineering case of a hydropower station on Jinshajiang River,a rock mass quality classification model is established with crossing verification and grid search optimized support vector machine model,for which 7 parameters,i.e.rock uniaxial compressive strength(Rc),rock quality index(RQD),rock weathering degree,number of joint set(Jn),joint roughness coefficient(Jr),joint alteration coefficient(Ja),groundwater state,are selected as the input parameters for buildng the classification model,so as to make the quality classification for the complicated rock structures within the dam site.Through the comparison made between RMR(Rock Mass Rating)and BP neural network classification method,it is indicated that the support vector machine not only has a high nonlinear mapping capacity with a quite strong ability for recognizing the rock classification,but also has better accuracy and stability,thus can meet the relevant demand of the actual construction.关键词
支持向量机(SVM)/岩体质量分级/BP神经网络Key words
support vector machine(SVM)/rock mass quality classification/BP neural network分类
计算机与自动化引用本文复制引用
何云松,薛秋池,赵其华..基于改进向量机的岩体质量分级研究[J].水利水电技术,2017,48(1):133-138,6.基金项目
国家自然科学基金(41272333) （41272333）
