ABSTRACT
Foundation scour poses a huge potential threat to bridge safety. A scour detection method of bridge pile foundation based on pier vibration mode excited by vehicle braking action was proposed and verified by numerical simulation and a laboratory test. Firstly, the horizontal vibration of bridge was excited by vehicle braking action, and the vibration modes sensitive to scour were identified from vibration signal. Secondly, the frequency change rate FCR and the flexibility-based displacement were defined as indicators to determine the scour location. Finally, with pier mode as input and scour state as output, a BP neural network model optimized by particle swarm optimization algorithm was constructed to predict the scour depth. Numerical and experimental studies showed that the vehicle braking force can effectively stimulate the local modes of the pier, and the scour indicators were very sensitive to scour. The scour depth can be accurately predicted using a small number of training samples and several vibration measuring points on the pier. The proposed method is simple to operate and can be embedded in the regular bridge load test to evaluate the bridge substructure synchronously.
