Research on Oil Well Fault Diagnosis Based on Multi-Dimensional Coupled Feature Extraction

With the development of artificial neural network and intelligent computing, the feature value extraction of fault samples in the oil well fault diagnosis of intelligent oil field is not complete at the present stage, so in order to detect the oil well fault condition timely and accurately, In this paper, a multi-dimensional coupled eigenvalue extraction method based on Continuous Wavelet transform-Fast Fourier Transform (CWT-FFT) is proposed to improve the accuracy of fault sample eigenvalue extraction.To improve the accuracy of fault sample eigenvalue extraction.Multi-dimensional coupled eigenvalue extraction is to carry out three-dimensional joint feature extraction through synchronous calibration of one-dimensional data with twodimensional coefficient matrix and three-dimensional energy view to retain transient and steady-state information and avoid the loss of single dimensional eigenvalue.Firstly, the existing fault data is preprocessed to obtain the standard data. Then, the eigenvalue is extracted by CWT-FFT model to generate eigenvalue matrix data, which is verified by the trained Convolutional Neural Networks (CNN) classification model. The results show that this method can improve the accuracy of fault classification of oil well significantly.