基于Fisher判别准则的公路隧道围岩分级评价

该文借鉴Fisher判别理论,建立公路隧道围岩评价模型,以期取得较好的效果。该项研究借助Fisher判别明显的统计优势,综合国内外大量岩体围岩分级指标资料及工程实际,建立了公路隧道围岩识别的Fisher判别分析模型,选用隧道洞室围岩岩石单轴饱和抗压强度σc、节理平均间距d、岩石质量指标RQD、地下水情况W等方面的综合指标作为围岩类别判定的评价因子,以21组公路隧道围岩实测样本数据进行学习训练和检验,并建立了相应的线性判别函数对13组待判隧道围岩类别情况进行了预测,判别结果与实际一致。研究结果表明,FDA模型回判估计误判率为0,交叉确认估计误判率为4.8%,判别性能稳健可靠,对现有评价公路隧道围岩分级方法进行了有效的验证和补充。

Assessment on Classification of Surrounding Rock Mass in Highway Tunnel Based on Fisher DiscriminantCriterion

Based on Fisher discriminant theory, the article sets up the surrounding rock mass assessment model of highway tunnel for achieving the better effect. With the statistically significant advantage of Fisher discriminant, and the large of surrounding rock mass classification index data and engineering practices at home and abroad, a highway tunnel rock identification of Fisher discriminant analysis model is established. The comprehensive indexes in the aspects of the uniaxial saturation compressive strength σc, average spacing of joints d, values of rock quality RQD and groundwater condition W of surrounding rock mass of tunnel are selected as the assessment factors to determine the classification of surrounding rock. The measured sample data of the surrounding rocks in 21 groups of highway tunnel are used for the training and test inspection, and the relevant linear discriminant function is also set up to forecast the classifications of 13 groups of tunnel surrounding rocks to be determined. The discriminant results are consistent with the actual discrimination. The study results show that the back-misjudgment ratio of FDA model is 0, and the misjudgment ratio of cross-validation method is 4.8%. The discrimination performance is stable and reliable. The article efficiently verifies and adds the present assessment method for the classification of surrounding rock mass in highway tunnel.