Fault Diagnosis Method for Wind Turbine Current Signal Based on Relation Networks and Margin Learning

The gearbox plays a crucial role in wind turbines, as they often operate in harsh environments for extended periods, resulting in a high failure rate for gearboxes. Therefore, fault diagnosis of gearboxes is essential. However, predicting these faults is challenging, especially when fault data is limited. To address this challenge, this study proposes a relational marginal learning neural network model. The model uses the current signals of the gearbox as input and integrates two different branches—a relational module and a fully connected layer—to learn the features of the samples. The relational module is designed to maximize intra-class compactness, while the fully connected layer aims to enhance inter-class separability. Experiments were carried out on a real wind turbine to verify the proposed method. The results indicate that the model can achieve a high accuracy of fault classification for gearbox faults. Ablation and comparison experiments were both conducted, which further validates the superiority of the proposed method.