When bridge towers are subjected to strong winds, they exhibit significant displacements. This displacement change can serve as an important indicator for structural performance warning. The displacement and acceleration collected in real time by the intelligent bridge monitoring system are disturbed by various noises, resulting in missed alarms in the monitoring system and causing huge economic losses. This study employs the fractional-order Butterworth lowpass filter method, eliminating the maximum value method, triple standard deviation method, etc. for preprocessing abnormal monitoring data characterized by missing values and outlier points. A fractional-order multi-rate Kalman fusion is proposed to process and model the correlation of structural displacement and acceleration data, and the simulated data and measured data are analyzed and verified respectively. Spectral analysis confirmed that by effectively fusing the low-frequency GPS signal with the high-frequency accelerometer signal, the fractional-order multi-rate Kalman fusion displacement measurement has a relatively high accuracy. Displacements obtained by the fractional-order multi-rate Kalman fusion method are adopted for correlation modeling, and residuals generated from this fractional-order fusion modeling are used for structural performance warning testing. The effectiveness of this structural performance warning is quantitatively validated through statistical assessment of warning accuracy.
Wang et al. (2026) studied this question.