Classification of Partial Discharge in Electric Aircraft based on Short-Term Behavior of Insulation Systems

Abstract

warming crisis has started a movement toward the reduction of emissions and fossil-fuel dependency. A major potential resides in the transportation sector which has led to the proliferation of electric vehicles. Similarly, significant progress has been made toward electrification in the aviation industry. Unlike electric cars, electrification in the commercial aircraft industry still looks far. One of the major bottlenecks in this path is the reliability of electrical equipment in aeronautical applications. The harsh environmental conditions along with the higher electric tension imposed by the high-power density equipment (which are necessary for the aviation industry) can accelerate the aging of insulation systems. The insulation system is regarded as the heart of the electrical equipment and its failure results in the breakdown of the system. A major deterioration mechanism in the insulation system is partial discharge (PD). In this study, we develop a framework for the condition monitoring of insulation systems at high altitudes based on deep learning. Different cases of corona discharge are tested based on standard IEC 60270 as potential sources of threat to the electrical systems. The measured signals are the input of the Dielectric Online Condition Monitoring System (DOCMS) that preprocesses the data, converts them into phase-resolved PD (PRPD) images, and classifies them based on their source type using EfficientNet. Then, DOCMS updates the system engineer about the status of electrical equipment in case of unsafe operation. The results demonstrate the high accuracy and fastness of the proposed approach to identify potential threats to the health of the insulation systems.

Authors

• Moein Borghei Avalanche Energy
• M. Ghassemi

Keywords

• Environmental Science
• Engineering
• Physics