Nontechnical Loss Detection With Duffing–Holmes Self-Synchronization Dynamic Errors and 1D CNN-Based Multilayer Classifier in a Smart Grid

A Comprehensive Investigation into the Application of Convolutional Neural Networks (ConvNet/CNN) in Smart Grids

“…For the evaluation of manual operation correctness in smart grid virtual reality training systems, an approach based on the vectorized spatiotemporal graph convolutional is effective [104]. In comparison to the conventional multilayer classifier, the combined Duffing-Holmes-based quantizer and one-dimensional CNN-based multilayer classifier exhibit promising performance and efficiency in feature extraction, training and recall processing, and accurate classification [105]. Autoencoder-bidirectional GRU model performs better in detecting the non-technical losses in SG and preventing it from energy theft [106].…”