Research On Damage Detection Techniques Based On Coupled Vehicle-bridge Vibration Responses

This research is sponsored by the National Natural Science Foundation of China (grant No.51078029).Bridge damage detection methods often need to interrupt traffic and close bridges, so the tests are often limited to be taken in the early morning to avoid the impact ontraffics, and it is difficult to interrupt traffic in some very busy lines. Therefore, how to detect the bridge damage without affecting the normal operating of bridges becomes one of the most concerned research topics.The bridge damage detection method based on the bridge response under working condition does not need to disrupt traffic, and is conducive to long-term health monitoring. The main research work and results are as follows:1. A technique based on the response of vehicle rather than that of bridge is used to identify the bridge frequency. A three-span continuous bridge and a1:10model train are designed to verify this technique. It is confirmed that the bridge frequency can be easily identified from the response spectrum of the vehicle.2. The sensitivity of the responses of bridges and vehicles to the bridge damage is studied. The responses of a few measure points of the bridge and vehicle are used to detect the structure damage and the change of the support stiffness by the sensitivity method. Moreover, the influence of different parameters on the detect results is studied. Besides, when the number of the bridge element is too many, some special methods are presented, such as elements reduction methods and results visualization.3. The concept of bridge vibration energy density is defined, and a new method is proposed to detect bridge damage using the vibration energy density sensitivity. This method has advantages in multiple measuring points, improving the defects of response sensitivity method which does not make full use of the connections of each measuring point and has a large amount of calculation, and can greatly inhibit the high noise of individual measuring point. A concept of windowed energy density is proposed further, which can get more accurate result and smaller calculating amount.4. A new method is proposed to detect bridge damage using the response cross-correlation function sensitivity, in which the acceleration responses of two measuring points on the bridge are recorded, then the changes of the cross-correlation function of these two responses are compared before and after bridge is damaged, and the sensitivity method is used to identify the bridge damage. The response cross-correlation function sensitivity method can increase the accuracy of identified results, which has advantages of simple operation, fast convergence rate, and high ability of inhibiting the noise of measuring point. It can also offer good results in identification of multiple damaged locations in bridges.5. Two wavelet sensitivity methods are studied to detect bridge damage using the discrete wavelet transformation and wavelet packet decomposition of the bridge responses, respectively. These two methods can get more accurate result and smaller calculating amount. A technique based on continuous wavelet analysis of the responses of the bridge and vehicles is also proposed for bridge damage detection.6. A framework is presented to detect bridge damage with uncertain parameters of passing vehicles based on the dynamic response of bridge. By computing the sensitivities of the identified damage indices responses to the vehicle parameter, the mathematical expectations and standard deviations of the damage indices can be obtained. Furthermore, the probability estimate of the damage indices can be obtained.