Investigation On Structural Thermal Effect Of Single-Tower Cable-Stayed Bridge Structural Health Monitoring

Temperature effect is one of the important factors affecting the safety of bridge structures.The temperature effect mechanism of single-tower cable-stayed bridges is complex.Analyzing the temperature effect of bridge structures and revealing the mechanism of temperature effect on the deformation of single tower cable-stayed bridges are of great significance in the structural health monitoring of bridges.This article takes the Ningbo Waitan Bridge as the research object,uses the long-term monitoring data of the bridge,conducts a study on the temperature effect of a single-tower cable-stayed bridge,derives a simplified calculation method for the temperature deformation of the single-tower cable-stayed bridge structure,proposes a new method to eliminate the temperature time-lag effect,and establishes a correlation model for the monitoring data of structural deformation.On this basis,a warning method for abnormal deformation of the structure is proposed,and an early warning for abnormal deformation of the bridge structure is carried out.The main research work and conclusions of this article are as follows:（1）Through plane geometry and structural mechanics analysis,the temperature deformation characteristics of single tower cable-stayed bridges are studied,and the temperature induced deformation formula of single-tower cable-stayed bridges is derived.The mechanism of temperature effect on the structural deformation of single tower cable-stayed bridges is revealed.The formula is verified by long-term monitoring data of the Waitan Bridge.The results show that there is a significant linear correlation between the main girder deflection and the girder-end displacement of the single-tower cable-stayed bridge under temperature.The temperature induced the main girder deflection of the Waitan Bridge is 2.5times greater than the girder-end displacement.The calculation result of this formula has high accuracy.The measured value of the deflection change of the main girder caused by temperature is 0.274 mm·℃^-1,and the calculated value of the formula is 0.287 mm·℃^-1,with an error of less than 5%.The calculation formula for temperature induced deformation of single tower cable-stayed bridges proposed in this paper has a high accuracy and can provide a reference for the design analysis and state evaluation of similar bridge structures.（2）Based on the recognition of the linear correlation between bridge structural deformation and temperature deformation,a time domain translation method to eliminate the temperature time-lag effect is proposed.Taking the Waitan Bridge as the research object,the temperature time-lag effects of pylon tilt,main girder deflection and girder-end displacement were eliminated,and the elimination effects were compared with those of Fourier series method.The analysis results show that there is a significant time-lag between the pylon tilt,main girder deflection and girder-end displacement,and the ambient temperature,with a time lag of 2.8 hours,1.7 hours,and 1.5 hours,respectively.By using the time domain translation method to eliminate the temperature time-lag effect,the linear correlation between structural deformation and temperature is significantly improved.Compared with Fourier series method,the method in this paper can more effectively eliminate the temperature time-lag effect,and the implementation process is simpler.It can be widely applied to the correlation analysis of bridge structure temperature and deformation monitoring data.（3）Based on Long and Short Term Memory（LSTM）neural networks,a correlation model between bridge structural deformation and ambient temperature monitoring data was constructed,and the prediction accuracy of the model was compared with that of Gated Recurrent Unit（GRU）neural networks.The results show that there is a multidimensional time lag phenomenon between the girder-end displacement,main girder deflection,pylon tilt,and environmental temperature monitoring data,and the correlation is unclear.It is difficult to accurately estimate the complex mapping relationship between the structural deformation using traditional linear fitting and nonlinear fitting methods.By optimizing the input structure of the model,the prediction accuracy of the LSTM neural network model for single variable monitoring data was improved,resulting in a decrease in Root Mean Square Error（RMSE）from 0.18mm to 0.07mm and a decrease in Mean Absolute Error（MAE）from 0.15mm to0.06mm.LSTM neural network model eliminates the multi-dimensional time-delay phenomenon of structural deformation monitoring data,and it’s prediction accuracy is significantly better than GRU neural network.The prediction accuracy of the multivariate prediction model is somewhat lower than that of the univariate model,but it can simultaneously predict the girder-end displacement and main girder deflection monitoring data series,reducing the training cost of the model,and is suitable for rapid prediction of multivariate deformation data.The LSTM neural network model established in this paper can provide a basis for damage early warning of bridge structures.（4）Using the LSTM neural network model established previously,combined with the mean standard deviation control chart,a method for early warning of abnormal deformation of bridge structures is proposed.The method uses the difference between the prediction data of the LSTM neural network model and the real-time deformation monitoring data of the bridge structure as an early warning indicator for abnormal structural deformation.When the bridge structure has abnormal deformation,and the resulting structure deformation monitoring has outlier,the original balance point of the control chart is broken,and the data point sequence exceeds the control line.After restoring the normal state,the data point sequence returns to the control line of the control chart.This method can achieve dynamic early warning of abnormal deformation of the bridge structure.