A Novel Training and Testing Platform for HIF Artificial Intelligence Detection

Artificial intelligence (AI) is widely used in the research of high impedance fault (HIF) identification in the distribution network. There is a lack of accurate HIF model, rich HIF samples and efficient automation platform in AI classifier training and testing. The paper first proposes a HIF model selection method based on discrete Frecher distance, and then constructs the waveform proliferation method to enrich the training sample library, and an automated platform is developed to efficiently and accurately train and test AI HIF classifier.