Fault Diagnosis of Electric Submersible Pumps Using a Three‐Stage Multiscale Feature Transformation Combined with CNN–SVM

A convolutional neural network support vector machine (CNN–SVM) method based on multichannel feature fusion is used for progressive fault diagnosis of offshore oil and gas wells. The excellent classification performance CNN attributed to its ability extract representations from large amounts easily distinguishable data. However, the capability severely constrained by noisy small sample amount electric submersible pump data be studied in this article. First, 12 representative statistical features are extracted raw reduce noise. Then, mapping model designed migration learning. Finally, SVM instead softmax function adopt directly classification. Comparative experimental results show that accuracy using feature‐extracted samples better than directly. proposed CNN–SVM approach has best compared SVM, BPNN, CNN, BPNN–SVM, CNN–Attention, CNN–LSTM, CNN–LSTM–Attention, which implies manual extraction still an indispensable tool process.