基于混沌时间序列的桥墩沉降预测

Pier Settlement Prediction Based on Chaotic Time Series

跨海大桥因船体撞击、移动车辆荷载、河水冲刷等因素的作用,桥墩的沉降变形表现为非线性特征。因此,首先对桥墩的沉降时间序列求取延迟时间τ和嵌入维数m,并采用最大Lyapunov指数证明该时间序列具有混沌特性;然后根据求取的参数建立加权零阶局域预测模型和加权一阶局域预测模型分别对沉降时间序列进行预测。算例结果表明,加权一阶局域预测模型具有较高的预测精度,且混沌局域预测法不适合做长期预测,但可做短期预测。

The cross-sea bridge is characterized by non-linear characteristics due to the hull collision,moving vehicle load,river erosion and other factors. In this paper,the delay time and the embedding dimension are obtained for the settlement time series of the piers,and the maximum Lyapunov exponent is used to prove the chaotic characteristics of the time series. Then the weighted zero-order local prediction model and the weighted first order the local prediction model predicts the settling time series. The results show that the weighted first-order local prediction model has high prediction accuracy,and the chaos local prediction method is not suitable for long-term prediction and can be short-term.