Hurst-Exponent-Based Detection of High-Impedance DC Arc Events for 48-V Systems in Vehicles
Yousef Abdullah, Jamie Shaffer, Boxue Hu, Bailey Hall, Jin Wang, Amin Emrani, Babak Arfaei
Abstract
Expanding dc electrical distribution systems require improved detection and mitigation of dc arc events. The high heat resulting from dc arcing raises concerns on equipment and personnel safety. In this article, a new adaptive detection method for series dc arc faults is presented. The proposed detection method utilizes Hurst exponent estimation of the system current to detect arcing events with a high detection rate and strong immunity against sudden load changes, chaotic load profiles, and switching harmonics. To verify the proposed detection method, a test setup to generate dc arcs was built, and over one thousand tests were performed under various conditions. The tests were conducted under different environment temperatures, air gap lengths, loads, and electrode diameters. The effectiveness of the proposed method is demonstrated by the test results and validated with an arc detector prototype.