Applied Research On Defect Detection Of Bridge Pile Foundation By Improved Acoustic Transmission Method

As a cheap and high efficiency nondestructive testing,acoustics transmission method has been widely utilized in detection of bridge pile foundation.However,there are still some problem in acoustics transmission method,which lead that acoustics transmission method cannot meet the demand of bridge detection.Therefore,an improved method was putted forward in this research.This method based on acoustics transmission method showed little influence by construction environment and material characteristics.Meanwhile,this method could recognize bridge defects automatically.In this research,waveform characters of different defects were analyzed in the view of principles of sound propagation.Then basic approaches and disadvantages of acoustics transmission method was discussed.Based on these disadvantages,an improved method based on acoustics transmission method was putted forward.This method includes three steps:(1)collect acoustic signal,(2)median filtering,(3)recognize defects based on convolutional neural networks.Finally,this improved method was utilized in detection of Ba-Wang bridge.The analysis of waveform characters from different defects showed that different defects had obviously different waveform characters.The detection results of Ba-Wang bridge showed that median filtering could weaken the influence of concrete.Meanwhile,convolutional neural networks could recognize different defects based on waveforms automatically and efficiently.