论文部分内容阅读
复杂工程水文地质条件下,为提高公路隧道施工和运营的整体安全性能,避免人员和财产损失,基于粒子群算法优化的支持向量机(PSO-SVM)判别理论,构建山岭公路隧道水害的危险性的分级评价模型。将隧道水害危险性分为4个等级作为SVM分类器的4个标签,依据“物质-构造-自然-人工”的事物发展逻辑,选取围岩岩性、岩体质量分级、构造断裂带类型、隧道围岩水系连通性、降水量、汇水面积、地下水高程差、防排水措施、爆破振动、隧道施工分级共10项指标作为SVM的判别指标。收集整理国内典型的20组公路隧道的指标数据作为模型的训练样本,将训练后的模型应用于济南二环高速公路项目上在建的6条隧道水害危险性分级中。研究结果表明:构建的PSO-SVM的分级模型准确率高,分级效果合理有效,及时提出了相应的公路隧道水害的预防和治理措施,有效地减少了隧道水害的发生,为类似隧道工程水害危险性分级提供参考和借鉴。
Under the condition of complex engineering hydrogeology, in order to improve the overall safety performance of highway tunnel construction and operation and avoid the loss of personnel and property, the PSO-SVM discriminant theory is adopted to construct the risk of water hazard of mountain highway tunnel Grading evaluation model. According to the logic of things of “material-structure-nature-man-made”, four criteria of tunnel water hazard are classified into 4 grades as the SVM classifier, and the surrounding rock lithology and rock mass classification are selected to construct the fault zone 10 types of indexes, such as the connectivity of surrounding rock system, precipitation, catchment area, groundwater elevation difference, drainage and drainage measures, blasting vibration and tunnel construction grading, were selected as discriminant indexes of SVM. Collect and sort the typical domestic 20 sets of highway tunnel index data as the training samples of the model and apply the trained model to the 6 tunnel water hazard classification under construction of Jinan Second Ring Expressway Project. The results show that the classification model of PSO-SVM has high accuracy, reasonable and effective grading effect and put forward corresponding prevention and treatment measures for water damage of highway tunnel in time, effectively reducing the occurrence of tunnel water hazard and providing a scientific basis for similar water hazard of tunnel project Sex grading provide reference and reference.