Defect Recognition and Classification Techniques for Multi-layer Tubular Structures in Oil and Gas Wells by Using Pulsed Eddy Current Testing

Pulsed eddy current (PEC) method has attracted researchers’ attention in the detection of multi-layer tubular structure defects in oil and gas wells because of its advantages of non-destructive testing, not being affected by non-ferromagnetic interference. In this paper, the detection experiment of the casing with prefabricated defects is carried out, and the PEC signal analysis method for defect recognition and classification of multi-layer tubular structures is studied. To deal with the problem that it is difficult to extract time-domain features, a modified principal component analysis method is proposed. Combining the statistical features extracted by the method and frequency-domain features, the combined features are obtained, and the random forest algorithm is introduced to classify the defects. This paper demonstrates an effective method for the detection of multi-layer tubular structure defects in oil and gas wells.