论文部分内容阅读
为准确预测巷道围岩稳定性类别,提出了基于网格搜索法(GSM)优化支持向量机(SVM)的巷道围岩稳定性预测模型。选取22组巷道围岩数据作为学习样本,以水平地应力与巷道夹角、顶板岩性、水的影响和巷道断面积4个指标作为模型输入,巷道围岩稳定程度作为模型输出,同时为增强模型的泛化性能和预测精度,采用改进的网格搜索方法优化支持向量机参数,最终构建基于GSM-SVM的巷道围岩稳定性预测模型。然后运用该模型对8组巷道围岩数据进行预测,并同BP神经网络模型的结果进行对比。结果表明,GSM-SVM模型的预测结果与实际结果吻合,正确率达98%,具有比BP神经网络模型更高的精度。
In order to accurately predict the category of surrounding rock stability, a prediction model of roadway surrounding rock stability based on grid search method (GSM) optimized support vector machine (SVM) is proposed. The data of surrounding rock of 22 groups of tunnels are selected as learning samples. Four indicators of horizontal stress, angle of roof, lithology of roof, influence of water and cross-sectional area of roadway are taken as model input. The stability of surrounding rock of roadway is taken as model output, The generalization performance and prediction accuracy of the model are optimized. The improved grid search method is used to optimize the parameters of SVM. Finally, the stability prediction model of surrounding rock mass based on GSM-SVM is constructed. Then, the model is used to predict the surrounding rock data of 8 groups of roadways and compared with the results of BP neural network model. The results show that the prediction results of GSM-SVM model are in good agreement with the actual results, and the correctness rate is 98%, which has higher precision than BP neural network model.