Nondestructive Testing Technology of Construction Machinery Crack of Bridge Prefabricated Piles
In response to the challenge of detecting structural cracks in bridge precast pile construction machinery,this paper designs a nondestructive testing system based on eddy current sensor arrays and Convolutional Neural Networks(CNN).The system employs a high-density eddy current array for comprehensive information collection,utilizes wavelet decomposition and Hilbert-Huang trans-form to extract crack signal features,and achieves crack type identification and severity assessment through an optimized CNN model.Experimental results demonstrate that the proposed nondestructive testing method can accurately and efficiently detect both macro and micro cracks in bridge precast pile construction machinery,with a crack detection rate of 99.7%,an identification accuracy of 97.2%,and a detection speed increased to 12.7 meters per minute,providing a powerful tool to ensure the safe operation of construction ma-chinery.
NETL
NSTL
万方数据
