A Novel Multi-Source Fault Diagnosis Strategy Based on Knowledge and Data Dual-Drive for a Planetary Gearbox

Traditional fault diagnosis methods face challenges such as the insufficient utilization of fault information and an imbalance between classification accuracy and computation. To address these issues, a novel multi-source fault diagnosis strategy based on a knowledge and data dual-drive algorithm is proposed. Firstly, a multi-source information correlation matrix (MICM) is designed to enhance the expression of fault information by combining information among time domain, frequency domain and channel correlation features. Then, kernel principal component analysis (KPCA) is used for dimensionality reduction in the MICM. Finally, a novel classifier based on Softmax logical regression (SLR) and K nearest neighbor (KNN) is proposed, where SLR provides an initial pre-classification and KNN is used to achieve accurate classification with less computation. Moreover, the latest planetary gearbox dataset of the wind turbine in a physical experiment is utilized to verify the effectiveness of the proposed MICM-SLR-KNN algorithm, and the experimental results demonstrate the superiority of the algorithm in comparison with other methods.