基于RBF神经网络的沙土液化预测研究

Prediction of Liquefaction of Sand Soil Based on RBF Neural Networks

简要分析RBF网络的结构特点及最近邻聚类学习算法,以大量粉土地基实测数据为学习训练样本及预测样本,建立了预测模型。讨论了基于RBF神经网络技术的沙土液化预测分析方法及其有效性。研究表明,用RBF神经网络方法进行沙土地震液化预测是可行的。

The structure features of radial basis function and the nearest neighhor-elustering algorithm have been discussed in detail. Based on a large amount of sand soil data, a model for prediction is set up and practical effectiveness of the theory RBF neural network is discussed. The study shows that the method is feasible to predict the liquefaction of sand soil.