| Bridges are the throat of highway and the key node of transportation interconnection.To ensure the safety of long bridges,bridge health monitoring system is widely used.As the main transmission element of cable-bearing bridges to transfer bridge deck load,the cable force value of cable-bearing bridges is an important monitoring index to evaluate the function and safety of bridges.In general,the cable force response collected by the health monitoring system is the result of the joint action of constant load,moving load,temperature and other factors,among which the temperature effect has a significant impact on the change of the cable force amplitude,which often masks the cable force response generated by the load to a certain extent.Therefore,separating the temperature effect in the cable force monitoring data and obtaining the real cable force caused by the load is of important research theoretical and engineering practical value for determining the real service condition of the bridge structure.In this paper,in response to the problems of poor accuracy and poor generalizability of the existing bridge response monitoring data temperature effect separation methods,a research on the bridge cable force monitoring data temperature effect separation method based on the combination of improved variational mode decomposition(IVMD)and K-L divergence(KLD)is carried out from mathematical statistics,correlation analysis,signal processing,and numerical simulation.The main study contents and conclusions are as follows:(1)The correlation between temperature and cable force at different time scales was analyzed.Firstly,the statistical analysis of the daily and annual temperature data of each section of the main beam was carried out,and the distribution pattern of the main beam temperature from the middle of the span to both sides of the bank and the standard value of temperature effect were obtained.The analysis results show that: the standard value of temperature effect decreases from the middle of the span to the two shore sides;the existence of daily and annual temperature effect components in the data spectrum of the cable force affected by temperature lays a feasible foundation for the application of the modal decomposition method to dissociate the temperature effect;the cable force at different locations is affected by temperature to different degrees,and the cable force near the bridge tower is more affected by temperature.(2)A process framework for the separation method of the temperature effect of the cable force is developed.First,the modal center frequency coefficient of variation and signal-to-noise ratio values are used to optimize the improved variational modal decomposition method to decompose the cable force and obtain a series of modal components;next,the components are combined according to the similar alignment entropy values to avoid the loss of temperature effect components;then,the K-L scatter value between the components and the original cable force signal is calculated to identify the temperature effect component components;finally,the correlation coefficients,error statistics indexes are proposed The separation effect evaluation is carried out.The results show that the constructed IVMD-KLD model has complete theoretical support and can be used to realize the adaptive separation of temperature effects in the cable force data.(3)The validation of the temperature effect separation effect of the numerical simulation signal of the cable force based on IVMD-KLD was carried out.Firstly,based on the standard values of temperature effects derived from the statistical analysis of temperature data,the temperature load conditions were transformed into the Midas Civil finite element model to obtain the simulation signal of the cable force;then,the IVMD-KLD method was used to separate the daily and annual temperature effects in the simulation signal;finally,a comparison analysis with different modal decomposition methods was conducted.The results show that the proposed method can effectively separate the temperature effect in the simulated cable force signal,and the correlation coefficients of the signal waveforms before and after the separation of daily and annual temperature difference effects reach 0.9834 and 0.9903,which are 9.12% and 12.50%higher than those of EMD-KLD,and 10.48% and 11.13% higher than those of EEMD-KLD,respectively.improve the problems of modal mixing,endpoint effect,and pseudo-orthogonality.(4)A study on the separation of temperature effect of real bridge cable force monitoring data and its application was carried out.Firstly,the separation of the fast-change components was carried out by using Butterworth low-pass filter;then,the separation of the temperature effect of the real bridge cable force monitoring data was carried out based on IVMD-KLD;finally,the probability distribution form of the constant-load standard cable force was obtained.The results show that after the separation of the fast-change components,the Pearson correlation coefficients between the daily and annual temperature effects of the measured cable force separated by the proposed method and the daily and annual variations of temperature are-0.9571 and-0.9892,respectively,which verify the strong negative correlation between temperature and cable force,and improve 17.42% and 20.17% compared with EMD-KLD,and11.41% and 20.17% compared with EEMD-KLD,respectively.The IVMD-KLD method showed good applicability in the application of real bridge cable force data,with improvements of 17.42% and 20.17% compared to EMD-KLD and 11.41% and12.13% compared to EEMD-KLD,respectively.After separating the fast-change and temperature effects,the constant-load cable force at different locations obeys a normal distribution with different parameters under random disturbance factors,and the upper and lower limits of the rejection domain are obtained at a given significance level,so that the warning interval of the constant-load cable force can be classified. |
PDF Full Text Request