基于距离判别分析法的冲击地压预测研究

考虑影响冲击地压的矿山地质因素和开采技术因素,提出预测预报冲击地压危险性的距离判别分析方法。选用煤层开采深度、顶板岩性、地质构造复杂程度、煤层倾角、煤层厚度、开采方法、有无煤柱、炮采或综采8项指标作为距离判别分析模型的输入变量,并以工程实测数据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。研究结果表明,距离判别分析模型学习性能良好,预测精度高,回判估计的误判率为零,是冲击地压预测预报的一种有效而实用的方法。

Prediction of Rock Burst in Mining Based on Distance Discriminant Analysis Method

A new method to predict rock burst is presented by considering the effect of mining techniques and geological conditions on rock burst. Eight indexes including uniaxial compressive strength, p-wave velocity in rock, volumetric joint count, joint roughness coefficients, joint weathering coefficients and permeability coefficient are selected as the input variables, and thus a distance discriminant analysis model is established. The discriminant functions of this model are obtained by training a large set of study samples, and then applied to engineering. The results show that this model has an excellent performance, high prediction accuracy and can be used in practical engineering.