A hybrid intelligent approach for metal-loss defect depth prediction in oil and gas pipelines

Abstract

The defect assessment process in oil and gas pipelines consists of three stages: defect detection, defect dimension (i.e., defect depth and length) prediction, and defect severity level determination. In this paper, we propose an intelligent system approach for defect prediction in oil and gas pipelines. The proposed technique is based on the magnetic flux leakage (MFL) technology widely used in pipeline monitoring systems. In the first stage, the MFL signals are analyzed using the Wavelet transform technique to detect anymetal-loss defect in the targeted pipeline. In case of defect existence, an adaptive neuro-fuzzy inference system is utilized to predict the defect depth. Depth-related features are first extracted from the MFL signals, and then used to train the neural network to tune the parameters of the membership functions of the fuzzy inference system. To further improve the accuracy of the defect depth, predicted by the proposed model, highly-discriminant features are then selected by using the weight-based support vector machine (SVM). Experimental work shows that the proposed technique yields promising results, compared with those achieved by some service providers.