Prediction of Bridge Monitoring Data and Time-Variant Reliability Assessment Based on Kalman Filtering

Effectively using the massive data of bridge health monitoring and controlling the safety status of bridge structure in real time is the research focus in the field of bridge health monitoring. In order to process the monitoring data online and evaluate the time-varying reliability of the bridge structure in real time, this paper proposes a dynamic prediction model based on Kalman filter to predict the daily maximum stress monitoring data during the service of the bridge by using an engineering example of a long-span cable-stayed bridge. Finally, the time-varying reliability index of the structure is established by using the prediction data, and the failure probability of the structure is obtained, and the online real-time dynamic prediction of the structural reliability is realized. The results show that the predicted value obtained by the dynamic prediction model based on Kalman filter has a high degree of fitting with the original data, and can accurately reflect the change trend and range of the original data. The convergence speed of prediction error variance is fast, and finally tends to 0. With the increase of the number of iterations, the absolute value of the relative error tends to decrease gradually, and the prediction accuracy of the model is higher. The time-varying reliability index and failure probability of the structure calculated by the predicted value based on the prediction model are highly fitted with the reliability index and failure probability based on the original data, which can reflect the change trend and range. In summary, the dynamic prediction model based on Kalman filter can still perform well in the case of less initial data, and realize real-time dynamic and long-term prediction of monitoring data. The time-invariant reliability of the structure can only be evaluated by offline data. In this paper, the dynamic prediction model based on Kalman filter is used to evaluate the time-varying reliability of the structure, which can realize the online real-time evaluation of the structural reliability and control the state of the structure or component in real time. It has important engineering significance.