A Library of Electrical Arc Signatures for the Development and Testing of PV Arc Faults Detection Techniques

Arc faults are a major cause of photovoltaic (PV) system failures, often leading to electrical fires due to equipment malfunctions or poor installations. Consequently, arc-fault detectors, now mandated in many jurisdictions, are essential for PV safety. Developing effective detectors necessitates a deep understanding of the chaotic electrical characteristics of PV arcs, which require analyzing a large sample of arc voltage and current waveforms. However, a significant obstacle is the scarcity of PV arc data in public domain literature. This paper presents an automated mechatronics testbed, developed using the pull-apart method, that enables repeatable arc generation studies. This testbed helps build a library of arc voltage and current waveforms, which aids in characterizing the electrical signatures of PV arcs and supports real-time testing of detectors. Various factors hypothesized to impact these electrical signatures, such as electrode geometry and material, electrical settings, and gap separation, are explored. Through extensive testing, a total of 2,352 arcs are generated by the testbed, with the data being collected and stored in a public repository. This curated library, made available to the public and is available on IEEE DataPort™, titled “Photovoltaic DC Arc Library”, supports further research and the development of effective PV arc fault detectors.