论文部分内容阅读
为对土石材料修筑的海堤运行状态实施有效分析和预测,在因果关系分析基础上,选取前期潮位因子、积分型降雨因子和时效因子,以径向基函数(RBF)神经网络为建模工具,结合实测序列特点,采用模糊C均值聚类算法比较确定计算中心,建立海堤安全监控RBF模型,实现海堤状态量的预测;在对模型误差序列的大小、趋势和分布特征分析基础上,提出基于置信度的预测效果的假设检验方法,并在给定置信水平下对不同预测时长的稳定性予以比较;以实例建立模型并对其训练及预测效果加以分析判别。
In order to carry out effective analysis and prediction of the operating status of the sea wall built by earth and rock materials, based on the analysis of the causation, the tide level factor, integral rainfall factor and aging factor of the early stage are selected. The radial basis function (RBF) neural network is used as the modeling tool , Combined with the characteristics of the measured sequence, the fuzzy C-means clustering algorithm is used to compare and determine the calculation center, and the seawall safety monitoring RBF model is established to predict the state quantity of the seawall. Based on the analysis of the size, trend and distribution characteristics of the model error sequence, This paper proposes a hypothesis testing method based on the confidence of the forecasting effect and compares the stability of different forecasting time under the given confidence level. The model is established and the training and forecasting results are analyzed and discriminated.