论文部分内容阅读
以页岩为研究对象,分别采用多元线性回归(MLR)及最小二乘支持向量机(LS-SVM)建立了页岩的单轴抗压强度及抗拉强度预测模型,考虑的间接指标包括:岩石密度、点荷载强度及纵波波速,并对上述两种预测模型进行了性能检验及比较。结果表明:页岩强度与密度、点荷载强度、纵波波速呈较好的线性关系,相关系数均大于0.89;MLR和LS-SVM方法均可得到较高精度的强度值,但单轴抗压强度的预测精度比抗拉强度高,更适合于抗压强度的预测。两类模型在预测岩石单轴抗压强度时效果相当,但LS-SVM方法更适合于抗拉强度的预测。
Taking shale as the research object, the uniaxial compressive strength and tensile strength prediction model of shale were established by MLR and LS-SVM, respectively. The indirect indexes considered include: Rock density, point load strength and longitudinal wave velocity. The performance tests and comparisons of the above two models were carried out. The results show that the shale strength and density, the point load strength and the longitudinal wave velocity have a good linear relationship, the correlation coefficients are all greater than 0.89. Both the MLR and LS-SVM methods can obtain higher precision intensity values, but the uniaxial compressive strength The prediction accuracy is higher than the tensile strength, more suitable for the prediction of compressive strength. The two models are quite effective in predicting uniaxial compressive strength of rock, but the LS-SVM method is more suitable for the prediction of tensile strength.