论文部分内容阅读
为有效预测地下工程岩爆的发生及烈度,结合地下工程岩爆的特点,分析岩爆影响因素及相关判别依据,选取围岩最大切向应力σ与岩石抗压强度σc之比σ/σc、岩石抗压强度σc与岩石抗拉强度σt之比σc/σt以及弹性能量指数Wet为判别因子,引入在线极限学习机理论,建立了岩爆预测的OS-ELM判别模型。以搜集到的国内外15组工程岩爆数据进行训练建模,训练完成后将样本数据做输出预测,得到模型的预测精度达97.98%,并与SVM、BP模型进行对比分析,结果表明,OS-ELM模型精度优于SVM和BP模型。利用该模型对国内两处隧道岩爆情况进行预测,结果与实际情况基本相符。研究表明,OSELM判别模型在岩爆烈度分级上具有良好的适用性和有效性。
In order to effectively predict the occurrence and intensity of rockburst in underground engineering and to analyze the characteristics of rockburst in underground engineering, the influencing factors of rockburst and the basis of discrimination are analyzed. The ratio of the maximum tangential stress σ to the compressive strength σc of rock is selected as σ / σc, the ratio of rock compressive strength σc to rock tensile strength σt, σc / σt, and elastic energy index Wet as discriminant factors, an on-line limit learning machine theory is introduced to establish the OS-ELM discriminant model of rockburst prediction. The 15 sets of engineering rockburst data collected at home and abroad were trained and modeled. After the training was completed, the output of the sample data was predicted. The prediction accuracy of the model was 97.98%, and compared with SVM and BP models. The results showed that OS The accuracy of ELM model is better than that of SVM and BP model. The model is used to predict the rock burst in two tunnels in China, and the result is in good agreement with the actual situation. The research shows that the OSELM discriminant model has good applicability and validity in rockburst intensity classification.