Visual Inspection of Fault Type and Zone Prediction in Electrical Grids using Interpretable Spectrogram-based CNN Modeling

In electrical grids, fault diagnosis (fault type and fault location classifica-tions) are critical due to their economic and important implications. Numerous smart grid applications have embraced data-driven methodologies. While the majority of the work in this topic has been on increasing the predicted accuracy of machine-learning model for fault diagnosis, one important aspect that has received less attention is the interpretability of these systems.We advocate for a complementary perspective. To represent faulty sig-nals, we propose a spectrogram-convolutional neural network based represen-tation of the electrical signals where pre-trained models such as GoogleNet and SqueezeNet are trivially used. We then perform multiple fault classifica-tion tasks and offer a visual interpretation of the collected findings. The sug-gested approach makes the model more transparent through the use of Gradient -weighted Class Activation Mapping (Grad-CAM), which visualizes regions in the input spectrogram that are more relevant for predictions, assisting the end-user in the understanding and interpreting the results. We explore the merits of the suggested technique in terms of increasing the transparency of the black-box ma-chine learning system, which is a critical requirement for designing modernized smart grids.

Automated summary

In the field of electrical grids, fault diagnosis is critical but often lacks interpretability. This paper proposes a spectrogram-convolutional neural network approach for representing electrical signals, using pre-trained models for multiple fault classification tasks and providing visual interpretation of the results to enhance transparency.