Classification of defective magnetic flux leakage inspection of drill pipe based on minimum entropy deconvolution and SVM-AdaBoost

Due to the complex working environment and corrosive substances in the drilling fluid, drill pipes will incur various types of defect on the surface during long-term use. Timely and accurate identification of defects in drill pipes not only saves costs associated with drilling but also effectively avoids triggering downhole accidents. In this paper, a magnetic flux leakage (MFL) inspection method combining the SVM-AdaBoost algorithm and the minimum entropy deconvolution (MED) method is proposed, aiming at the effective identification and classification of five different defect types in drill pipes. The study firstly utilises the minimum entropy deconvolution method to preprocess the magnetic flux leakage signal from the drill pipe in order to enhance the identifiability of any defect features in the signal and reduce noise interference. Secondly, the SVM-AdaBoost algorithm is used to extract and classify the features of the processed signals to improve the accuracy and robustness of defect recognition. The results, which are based on a combination of experiment and simulation, show that the method can effectively distinguish between different types of drill pipe defect with high classification accuracy, which is of great significance for improving the efficiency and reliability of drill pipe inspection.