Classification of Hydrate Blockage and Pipeline Leakage in Natural Gas Pipelines Based on EMD and SVM
Bin Yue, Xiaocen Wang, Zhigang Qu, Yang An, Shuo Jin, Liqun Wu, Likun Wang, Xiliang Yang
Abstract
In this paper, pipeline abnormal events (hydrate blockage and pipeline leakage) were detected by an active acoustic excitation method based on pulse compression. The positions of pipeline abnormal events were determined by time delay between emitted and reflected signals. Besides, in order to effectively distinguish the two detection signals, the empirical mode decomposition (EMD) method was used to obtain the intrinsic mode function (IMF) of the detection signal, and the normalized energy of all IMF components was used as eigenvector to input to the support vector machine (SVM) classifier for classification. The experiment results demonstrated that the trained models could accurately classify hydrate blockage and pipeline leakage after adjusting the classification threshold by the Youden index. The evaluation indicators including accuracy, precision, recall, specificity, and F1-score were 100% and area under curve (AUC) was 1 in testing set.