Dynavolt Tech (China)

Fault detection in wind turbines is critical to ensure reliable operation and reduce maintenance costs. Traditional methods usually rely on a single sensor signal, which makes it difficult to comprehensively capture complex fault characteristics. In this paper, a wind turbine fault classification method based on multi-sensor spatio-temporal attention network (MSSTANet) is proposed for the intelligent diagnosis of internal mechanical faults in wind turbines.The MSSTANet model is capable of extracting and classifying fault features more accurately by integrating the multi-sensor signals and effectively utilizing the complementary nature of multi-sensor signals. Experimental results show that the proposed method achieves an accuracy of 93.07% for recognizing five types of multi-sensor fault events, providing an efficient and reliable solution for wind turbine fault diagnosis.