基于混沌时间序列的桥梁沉降预测

Prediction of Bridge Settlement Based on Chaotic Time Series

混沌的离散情况常常表现为混沌时间序列,而混沌时间序列中蕴含着丰富的系统的动力学信息。本文基于某桥梁的实际观测的沉降时间序列用自相关法求取时间延迟t、用Cao方法求取嵌入维数获得相空间重构参数,然后用最大Lyapunov指数法进行时间序列的混沌特性识别,证明桥梁沉降运动系统具有混沌特性。最后分别使用加权一阶局域预测法、Volterra级数自适应预测法以及RBF神经网络预测模型进行预测,比较了几种方法的预测精度,得到RBF神经网络模型在短期预测中具有较好的性能。

Chaotic discrete time is often expressed as chaotic time series, and chaotic time series contain abundant dynamical information of the system. The settlement time series observed a bridge based on the autocorrelation method for time delay T, Cao method is used to calculate the embedding dimension for the phase space reconstruction parameters, chaos identification and time sequence using the maximum Lyapunov index method, prove the bridge settlement motion system with chaotic characteristics. Finally, using a weighted local prediction method, Volterra series method and RBF neural network prediction model, compare the prediction accuracy of several methods, RBF neural network model has better performance in short term prediction.