ArcSafe-60

The ArcSafe-60 dataset is a meticulously curated experimental collection designed for series arc fault detection using low-frequency (60 Hz) current signals in residential electrical environments. Data acquisition was conducted on a multi-branch AC arc fault test platform incorporating representative household appliances such as microwave ovens, rice cookers, and air conditioners under both single-branch and multi-branch parallel configurations. The dataset comprises synchronized current waveforms sampled at 60 Hz using high-precision Hall-effect sensors with comprehensive preprocessing including band-pass filtering and normalization to ensure signal integrity. ArcSafe-60 offers a diverse range of normal and fault conditions, facilitating robust training, validation, and cross-domain transfer learning for arc fault detection models. By addressing practical hardware constraints and providing realistic low-sampling-rate measurements, ArcSafe-60 serves as a valuable benchmark for developing lightweight, reliable, and real-time arc fault detection methods applicable to smart grid and IoT-enabled edge devices.