Detection Of Damaged Bridges Based On Indirect Measurement Method And Continuous Wavelet Transform

Due to the fast development of society and the continuous improvement of transportation systems,bridge monitoring has become an important problem to be solved.Traditionally,one can obtain the dynamic properties of a bridge directly through the sensors installed on the bridge,which is called the direct measurement method.However,the concept of indirect measurement method has its own way to acquire the dynamic properties of the bridge,which is made possible using the response recorded by a test vehicle during its travel over the bridge.In 2004,Professor YB Yang firstly proposed this new method.Compared with the conventional direct measurement method,not only is the indirect measurement method more economical and convenient,it also avoids the difficulties in maintenance or updates for sensors and the requirement in data storage capacity.Most importantly,the indirect measurement method has been studied and proven to be feasible by both theoretical and experimental researches.The purpose of this paper is aimed at proposing a damaged bridge model and studying the feasibility of damage detection based on the indirect measurement method.One rotational spring and one vertical spring are together used to simulate the reduction of bridge stiffness(i.e.,damage)of a beam,which is linked with two ordinary beam elements.The vehicle is still simplified as a one-degree-of-freedom sprung mass.The major process of bridge damage detection is as follows.First of all,record the vehicle acceleration during its passage over the bridge.Then,conduct continuous wavelet transform on the abstracted response.Third,locate the damage with the wavelet coefficients.Forth,determine the damage index according to the fitting equations which have been calculated for the previous on-site detection.Finally,estimate the damage level of the monitored bridge.In the paper,the wavelet function Db4 chosen as the basic function behaves well.In addition,we also conduct a parametric analysis on the moving vehicle.The results indicate that the heavier the vehicle,the more rigid the vehicle spring and the greater the driving velocity all contribute to the amplitude of wavelet coefficients,despite unsatisfactory detection sensitivity.Besides,one should pay more attention to the vehicle velocity,because the sampling number can vitally influence the accuracy of detection.Moreover,the damping of the moving vehicle is also beneficial to the detection accuracy.It is worth noting that increasing the scale of the wavelet function really increases its accuracy when white noise is added to vehicle responses,because the presence of high frequencies in white noise plays a bad role in detection of singularities or discontinuities.It is expected that the conclusions achieved above can be used as a reference for future studies on similar topics.