Research on Defect Type Identification of Thin-walled Stainless Steel Seamless Pipe Based on BP Neural Network

An impedance diagram analysis method of eddy current signal based on BP neural network was proposed, which can effectively identify the defect types of seamless thin-walled stainless steel tube. Firstly, the signal geometric features were extracted by Canny algorithm and Radon transform. Then, the dimension of defect feature vector reduces by fast principal component analysis (FPCA) to remove the redundant information and get the principal component vector of defect. The feature vector after dimension reduction is used to train BP neural network to realize the recognition and classification of defects in the case of large noise. Through the experimental detection of the defects of stainless steel seamless circular tubes with different specifications, it is shown that the method has good robustness and universality, and the accuracy rate is as high as 93.9%.