| Summary: | With the increasing use of renewable power technologies, integration of offshore wind power generation and continuous improvement to battery energy storage systems for electric transport applications in the automotive, shipping and aerospace sectors, there is an increased use of DC power networks at higher voltage and power levels. These higher power DC systems are at greater risk from DC arc failure, caused by the degradation of conductor insulation, component failure or low-quality manufacturing. These arc failures present a significant risk to the reliability and safety of these power networks, burning at thousands of degrees and potentially starting electrical fires that could be catastrophic for electric transport applications. As such, there is a necessity for the development of fast, accurate DC series arc detection algorithms for future transportation and power distribution applications.
In this work the Windowed Fractal Dimension (WFD) arc detection tech-
nique is developed and tested against empirical arc data captures generated
across multiple arc ignition types, numerous circuit loads and topologies,
and a range of different environmental conditions to assess its capability
for arc detection when compared to other methods in the literature. This
work demonstrates the ability of the WFD technique to successfully distinguish DC series arc failures from healthy circuit behaviour through a change in the fractal dimension of circuit voltage and current waveforms at arc ignition, presenting a novel DC series arc detection technique focusing on the fundamental fractal behaviour of the arc. Throughout, the WFD technique is shown to improve upon existing arc detection methodologies in the literature, demonstrating an faster 1.5 ms detection time, improved resilience to nuisance trip conditions, and displaying continued efficacy to detecting arc failure in a broader range of load and environmental conditions.
|
|---|