桥梁预制桩施工机械裂纹无损检测技术

Nondestructive Testing Technology of Construction Machinery Crack of Bridge Prefabricated Piles

针对桥梁预制桩施工机械结构裂纹检测难题,设计了一套基于电涡流传感器阵列和卷积神经网络(CNN)的无损检测系统。该系统采用高密度电涡流阵列实现全信息采集,利用小波分解和希尔伯特-黄变换提取裂纹信号特征,并通过优化的CNN模型实现裂纹类型识别和严重程度评估。试验结果表明,所提出的无损检测方法能够准确、高效地发现桥梁预制桩施工机械中的宏观和微观裂纹,裂纹检出率达99.7%,识别精度为97.2%,检测速度提升至12.7 m/min,为保障施工机械安全运行提供了有力工具。

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 transform 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 machinery.