A Parametric Model Approach to Arc Fault Detection for DC and AC Power Systems

Arc faults can arise in any electrical system as a consequence of loose connections, inadvertent damage during maintenance and/or insulation aging. In this paper, a new approach to the detection of arc faults is discussed. By using the proposed parametric models of the fault signals, an arc fault can be identified quickly and reliably, so that action can be taken to prevent the arc from escalating into a serious fire and causing loss of life and/or property. One model was obtained for DC power systems and another model was obtained for AC power systems. These proposed models were able to distinguish arc faults from normal operation of the various electrical loads that were tested