Series Arc Fault Detection by Modeling and Integral Regulated Residual Analysis

With the rapid development of renewable energy, photovoltaic systems, energy storage systems, and DC microgrids are widely used. However, arc faults can cause electrical fires and even compromise the safety and reliability of the system. In this paper, a modeling and integral-regulated-residual-analysis-based arc fault detection technique is developed for DC systems. The generalized state space average model of the DC system is developed under normal operation and arc fault conditions. The power electronic harmonics are considered in the modeling and analyzed to estimate the state variation of the system. The developed model can predict the arc current and can preliminarily determine the causes of variation. The integral regulated residual analysis method is proposed to calculate the intersection area between the model-predicted arc current and the actual arc current. The nominal values of the intersection area can be determined by experiments. Thus, an arc fault can be identified when the value of the intersection area reaches the nominal value. In this way, the modeling and integral regulated residual analysis algorithm is established and the arc detector is developed. The experimental results of 600 groups of tests verified that the technique can detect sustained and intermittent arc faults at different arc formation conditions with an accuracy of 99.5% within 80 ms despite the effects of line impedance, load transient and power electronics noise.