隧道岩爆预测的距离判别分析模型研究及应用

Model of Distance Discriminant Analysis for Rockburst Prediction in Tunnel Engineering and Its Application

将洞室最大切向应力和岩石单轴抗压强度的比值、岩石单轴抗压强度和岩石单轴抗拉强度的比值、弹性能量指数这3项影响隧道岩爆的主要指标作为判别因子,建立隧道岩爆预测的距离判别分析模型。以国内外重大隧道工程发生岩爆的实测资料作为学习样本进行训练,建立相应判别函数对待判样本进行预测。该模型在预测判别中综合考虑了影响隧道发生岩爆的多项因素,通过判别模型的学习功能获得岩爆与其各影响因素之间的复杂关系,排除了一般准则中建立判据时人为因素的影响。实例证明该模型利用回代估计法所得到的误判率为零,并具有较强的判别能力。

A distance discriminant analysis model of rockburst prediction in tunnel engineering is set up in this paper. Three indexes reflecting the rockburst of tunnel engineering, including the ratio of maximum tangent stress of excavation to rock uniaxial compressive strength, the ratio of rock uniaxial compressive strength to rock uniaxial tensile strength and the index of elasticity energy, were used as the discriminant gene of distance discriminant analysis model. With on-the-spot survey data of the rockburst occurred in some important tunnel engineering projects home and abroad as training samples, corresponding discriminant functions are, established to predict the samples to be discriminated. This model comprehensively takes into account many factors resulting in the occurrence of rockburst in tunnel for prediction and differentiation. The complex relations between rockburst and each influencing factor are obtained through the learning function of the discriminant model, which can eliminate the influence of man-made factors upon establishing criterion in general guideline. Practical tunnel engineering instance has proved that the rockburst prediction model of distance discriminant analysis has excellent discriminant performance and the resubstitution error rate is zero.