A Novel Compound Fault Diagnosis Method for Rotating Machinery Based on Dynamic Adaptive MWPE and Dual-Graph Regularization Strategy

Detection of compound fault in rotating machinery under complex operation environment is a challenge in fault diagnosis. Machine learning occupies an important position in the field of fault diagnosis due to its broad applicability and high efficiency, while feature extraction and feature selection are key aspects in the machine learning process. As a result, this paper aims to enhance the performance of compound fault diagnosis methods by improving these two aspects. Firstly, to address the non-linearity and non-stationarity of vibration signals under variable operation conditions, this paper proposes a dynamic adaptive multi-scale weighted permutation entropy (DAMWPE) method. In addition, this paper decomposes the vibration signals with improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and extracts the sensitive intrinsic modal functions’ (IMFs’) DAMWPE (SI-DAMWPE) as the initial feature vector, which more accurately reveals the intrinsic time-scale characteristics of the vibration signals. Secondly, to address the problem that most feature selection methods ignore the correlation between faults, this paper proposes a novel multi-label feature selection method called dual-graph regularization considering feature redundancy feature selection (DRFRFS). The method employs the feature and label graph regularization strategy to comprehensively capture the relationship between fault labels and features. Finally, the top-ranked features from the DRFRFS method are selected and fed into a multi-label k-nearest neighbor (MLKNN) classifier to complete the diagnosis task. By comparing six multi-label classification evaluation metrics for two rotating machinery cases, the results show that the proposed method possesses high accuracy and stability.