Realistic Time-Domain Simulations of Arc Faults in Low-Voltage Installations

Electric arc faults that appear in home installations pose a significant safety hazard that may result in an outbreak of fire. Such events can be prevented using an arc fault detection device (AFDD) that interrupts the electrical circuit upon detecting an arcing event. In this work, we propose a realistic model of an arc fault for electrical simulations of low-voltage installations. We model an arc fault as a circuit component whose parameters dynamically change to account for the stochastic behaviour of electric arcs. The model uses a black-box representation that consists of a set of mathematical equations that link current and voltage. The proposed model goes beyond the state of the art by dynamically adapting the arc parameters after each current zero crossing. The parameters and their statistical distributions are extracted from laboratory measurements using optimization methods. In addition to that, we implement a noise source within the arc model with power spectral density that follows 1/f law. Power spectral density is extracted from an actual noise measurement. Transient behaviour of the modelled arc fault is implemented in an analogue and mixed signal simulation engine based on VHDL-AMS. The developed component can be used as a part of more complete electric installation simulation.