DC distribution systems are becoming common in aircraft, ships, automobiles—but also in residential homes. A critical unique feature of dc distribution is the self-sustaining series arc fault. The focus in this work is to provide a reasonable model of this dc arc to assess its impact on dc systems.
Additionally, an imminent problem is the detection, localization, modeling, and simulation of arc faults. These arcs form when conductors (or connectors) fail, break, crack, or degrade. The arcs that form in dc, since self-sustained, and are sources of fire, skin burn, electrical shock, and asset damage.
The Center for Electromechanics operates a megawatt-level dc microgrid that operates connected to the grid or in island model. An example configuration is shown in Fig. 1. To understand the impact of dc arcs in dc microgrids, our microgrid has been faulted several times to capture significant fault data.
Microgrid at the Center for Electromechanics (one possible configuration)
Fig. 1 shows a photograph of an arc forming under accelerated conductor separation. This situation is occurs when conductors break and fall. Fig. 2 shows a sustained arc in the presence of slowly varying gap distance.
Fig 1 . DC arc fault under accelerated separation (800 V, 200 A)
Fig 2 . DC arc fault under steady separation (280 V, 50 A)
Modeling and Simulation
Because staging arc faults is destructive, it is important to simulate arc damages using a computer model before staging faults in practice. The Center for Electromechanics has developed a simple and accurate dc arc fault model. The model consists of a nonlinear resistance in series with a voltage source as shown in Fig. 4.
Fig 4 . Left: arc branch model showing voltages and currents terms. Right: how the arc branch model relates to two separating electrodes.
The model has been validated experimentally on our microgrid by staging three types of faults: constant-speed gap, fixed-distance gap, and accelerated gap. Comparisons of experimental and simulated faults are shown in Fig. 5-Fig. 7. These comparisons show how well the model can predict the arc’s instantaneous voltage, current, power, and energy.
Fig 5 . Case study 1: constant speed fault (top row: instantaneous voltage and current; bottom row: instantaneous power and cumulative energy).