Numerical signal processing methods for partial discharge detection in more electrical aircraft

In forecoming “more electrical aircrafts”, aircrafts will become more compact and lighter. Hydraulic, pneumatic or mechanical systems will be replaced by electrical ones thus leading to an increased demand of electric power. The network voltage is expected to increase far above Paschen's minimum, thus increasing the risk of partial discharges. Furthermore the use of inverter drives will be generalized and PD detection on square voltage powered equipments will become a mandatory challenge. The purpose of this study is to propose a method to denoise partial discharge signals embedded in the electromagnetic noise induced by switchings. This denoising is performed using the wavelet decomposition.