Research on Identification Method of Scour Depth for Bridge Based on ERA and SVM

Abstract

A new damage identification method for bridge scour was proposed, in the case that it was difficult to detect bridge scour depth applying testing equipment. Through integrative application of the eigensystem realization algorithm (ERA) and method of support vector machine (SVM), this method was used to identify the scour depths of bridge under conditions of ambient excitation. The following three steps are necessary for the application of this method to identify bridge scour. Firstly, a sample library about scour depth and upper structure vibration response was established using numerical methods and support vector machine method. Secondly, free response signal of bridge were extracted from random vibration signal of bridge upper structure using random decrement technique. Thirdly, based on above two steps, the bridge scour depth was identified using ERA and SVM. In the process of applying the method to identify bridge scour depth, the key is to find the sensitive points for scour depth of substructure using sample library and to gather the vibration response signal of accelerated velocity under conditions of ambient excitation. It was identified that the method has higher recognition accuracy and better robustness through experiments on a real bridge. The method provided a new way for identifying scour depth of bridges.