Fig 1.
The schematic of the monitoring equipment layout of the bridge.
Fig 2.
The dimension marking for the schematic diagram.
Fig 3.
Measured mean wind speed time series curve.
Fig 4.
The rose chart of mean wind direction characteristics.
Fig 5.
The scatter plot of the correlation between turbulence and mean wind speed for downwind turbulence intensity.
Fig 6.
The scatter plot of the correlation between turbulence and mean wind speed turbulence intensity in the crosswind direction.
Fig 7.
The RMS curve of measured vibration response acceleration in each direction.
Fig 8.
Distribution of partial autocorrelation functions for acceleration in different directions.
(a) Lag order of acceleration in the along the bridge direction; (b) Hysteresis order of acceleration in the transverse direction; (c) Lag order of vertical acceleration; (d) Torsional acceleration hysteresis order.
Fig 9.
Schematic of one-dimensional causal convolution.
Fig 10.
Principle schematic of dilated convolution.
Fig 11.
The structure of the residual connection block.
Fig 12.
The framework diagram of the bridge wind-vibration response prediction algorithm.
Fig 13.
The illustration of sliding window for median filtering.
Fig 14.
Correlation coefficients between wind characteristic parameters and RMS values of wind vibration response of bridges.
Table 1.
The range of hyperparameters and results.
Table 2.
Comparison of prediction errors of different.
Fig 15.
Acceleration prediction results of different models along the bridge acceleration.
Fig 16.
Acceleration prediction results of different models in transverse acceleration.
Fig 17.
Acceleration prediction results of different models in vertical acceleration.
Fig 18.
Acceleration prediction results of different models in torsional acceleration.
Table 3.
Comparison of training time for different algorithms (Unit: second).
Fig 19.
Comparison of training time.
Table 4.
Comparison of forecast time for different algorithms (Unit: second).
Table 5.
Comparison of inference speed for different algorithms (Unit: ms/patch).
Table 6.
The p-value result of paired t-tests for different prediction algorithms(Alpha level of α = 0.05).
Table 7.
Statistical characteristics of the four sets of wind data.
Fig 20.
Wind speed (left) and wind rose plots (right) were used for sensitivity analysis of raw data for group a.
Fig 21.
Wind speed (left) and wind rose plots (right) were used for sensitivity analysis of raw data for group b.
Fig 22.
Wind speed (left) and wind rose plots (right) were used for sensitivity analysis of raw data for group c.
Fig 23.
Wind speed (left) and wind rose plots (right) were used for sensitivity analysis of raw data for group d.
Table 8.
Sensitivity test results.