Inverter Fault Diagnosis based on GASF-Alexnet

To tackle the challenge of extracting fault features from Voltage Source Inverters (VSI) under dynamic conditions, this paper introduces a novel fault diagnosis method utilizing GASF-AlexNet. Initially, the Gramian Angular Field (GASF) transformation encodes time-series data into images, enhancing feature visualization and discriminability. Subsequently, an enhanced AlexNet deep learning architecture autonomously extracts critical fault features from these images. Moreover, this study employs an adaptive moment estimation (Adam) optimizer combined with a dynamic learning rate decay strategy to optimize the training process and enhance the model's diagnostic accuracy. Comparative analysis demonstrates that the proposed method exhibits superior fault identification performance and efficiency under various conditions compared to traditional techniques. Experimental results validate the effectiveness and superiority of this method in practical applications.