基于优化支持向量机的强夯有效加固深度研究

Study on the effective reinforcement depth of dynamic consolidation based on optimized support vector machine

强夯法运用重锤高速冲击实现工程场地不良土质的改良加固,在地基处理工程中应用广泛,而有效加固深度的标定是应用强夯实现地基处理时的核心环节。文章基于强夯理论与有效加固深度概念,提出了以统计学习原理为理论的支持向量机算法,回归预测了强夯有效加固深度;并应用遗传算法优化了支持向量机模型的参数,通过对工程样本数据的回归、预测,对比了预测精度与误差反向传播神经网络模型预测结果。结果表明:优化支持向量机模型能够在数据样本上实现准确的回归拟合;模型在测试样本数据的预测误差低、预测精度优于神经网络模型。

Dynamic consolidation,which uses the high-speed impaction of heavy hammer to achieve the improvement and reinforcement of the poor soil in the engineering site,is widely used in the engineering of foundation treatment.The calibration of effective reinforcement depth is the core link of foundation treatment by dynamic compaction.Based on the dynamic compaction theory and the concept of effective reinforcement depth,this paper puts forward the support vector machine algorithm by using statistical learning theory to make regression prediction for the effective reinforcement depth of dynamic compaction;meanwhile it applies a genetic algorithm to optimize the parameters of the support vector machine model.According to the regression and prediction of engineering sample data,it carries out a comparative study on the prediction results of back propagation neural network model with the prediction accuracy and error.The results show that the optimized support vector machine model can achieve regression fitting on the data samples accurately,and the model has low prediction error on the test sample data and owns better prediction accuracy than that of neural network model.