论文部分内容阅读
由于苏通大桥群桩基础的超长桩在不同断面的轴力沿桩长分布具有复杂非线性的特点,为了实现对不同高程断面的轴力进行预测,引入粒子群算法与蚁群算法对支持向量机进行优化,在考虑断面高程、潮位、水温与风速等影响因素的基础上,建立沿桩长基桩轴力分布的优化SVM预测模型。研究表明,PSO-SVM与ACO-SVM模型比传统SVM在预测精度、模型稳定性与泛化能力方面有更好的表现,在超长桩不同高程断面的轴力预测中,具有一定的工程应用价值。
Due to the complex non-linear distribution of axial forces along the pile length of the super-long piles with pile foundation in Sutong Bridge at different sections, in order to predict the axial forces of different elevation sections, the introduction of PSO and ant colony optimization Based on the factors such as elevation, tide level, water temperature and wind speed, the optimal SVM prediction model of axial load distribution along the pile foundation is established. The research shows that the PSO-SVM and ACO-SVM models perform better than the traditional SVM in predicting accuracy, stability and generalization ability of the model. The PSO-SVM and ACO-SVM models have some engineering applications in predicting axial force of different elevation sections value.