将Fisher判别分析(FDA)理论应用到深部硬岩岩爆预测中,以最大切向应力σθ、单轴抗压强度σc、单轴抗拉强度σt和弹性能量指数Wet作为影响岩爆的因素,建立以σθ/σc,σc/σt及Wet为判别因子的岩爆预测的FDA模型.以国内外15个深部岩体实例为训练样本进行训练,利用训练好的模型对冬瓜山硬岩矿山和某煤矿三水平西大巷深部开拓巷道岩爆进行预测.研究结果表明:利用FDA模型进行深部硬岩岩爆预测,结果与工程实际岩爆结果相符合,说明该方法在深部硬岩岩爆预测中具有良好的工程实用性.%Fisher discriminant analysis (FDA) model was established to predict the possibility and classification of rockburst. Four indexes, i.e., maximum tangential stresses σθ, uniaxial compressive strength σc, uniaxial tensile strength of rock σt, and elastic energy index of rock Wet, were chosen as the main control factors of rockburst in the analysis. Three discrimination factors including σθ/σc, σc/σt and Wet were regarded as the discriminant factors of the FDA model. Fifteen depth rock projects at home and abroad were taken as the training and testing samples. Rockbursts of two mines were predicted using this model for illustration. The prediction results were compared with the practical ones. The results show that the FDA model of deep hard rock rockburst prediction has excellent performance and low error rate. It is a useful way to predict deep hard rock rockburst.
展开▼
