Surface Discharge Detection and Severity Assessment of Printed Circuit Boards for Electric Aircraft Using Fluorescent Fiber

Abstract

More and more printed circuit boards (PCBs) will be used in electric aircraft to achieve higher power density of the on-board electric system, while PCBs in aircraft are more likely to generate partial discharges (PDs) due to external factors such as compact structure and low air pressure. Therefore, this article proposes a fluorescent fiber-based method for detecting and evaluating PDs on PCBs. The detection method is effectively immune against the presence of the electromagnetic, acoustic, and vibration interference. Based on the optical detection, the evolution regularities of PCB surface appearance changes, optical PRPD patterns, and optical pulses during the aging process of PCB under different air pressure are analyzed in this article. Then, 12 assessment features are extracted for the PD aging process, and the contribution of these 12 features to the severity assessment at different air pressures is obtained using the minimal-redundancy-maximal-relevance (mRMR) algorithm. Finally, different numbers of PD features are tested for PD severity assessment by the support vector machine (SVM) algorithm. The evaluation results show that the severity assessment method proposed in this article can achieve an assessment accuracy of at least 91.1% and up to 94.4% under all three air pressures, which has good application and guidance value.