Partial Discharge Detection and Assessment of Surface defects under Fast Rising Repetitive Square Voltage for MV Converter Application

Abstract

Because of higher power density and blocking voltage for Silicon Carbide (SiC) devices, careful insulation design is of much importance for future medium voltage power converters. Due to the insulation or mechanical coordination, surface defects (metal-insulation-air interface problem) inside the converters or high voltage PCB boards cannot be avoided. It is undeniable that the final design should be Partial Discharge (PD) free under normal operation with a safety margin. However, the traditional PD detection method could be ineffective or insensitive because of the serious disturbance of displacement current under fast rising repetitive square voltage, which leads to great difficulty in PD-based insulation assessment. Moreover, PD characteristics and mechanism of surface defects under PWM-like voltage, which contribute to guide the insulation design, also has not been understood clearly. For these purposes, this paper proposes a Silicon Photomultiplier (SiPM) based optical PD detection method firstly, which is suitable for offline or online PD detection and diagnosis under PWM-like voltage because of the advantages of small size, high gain, good sensitivity and high Electromagnetic Interference (EMI) immunity. Then, PD characteristics of artificial surface discharge model under PWM-like excitation are investigated and the relationship between the PD activities (average discharge magnitude and repetition rate) and the PWM-like excitation parameters (rising time, amplitude and frequency) has been explored. Finally, the proposed SiPM sensor is used for the insulation design and assessment of HV differential mode inductor. Experimental results indicate that space charge accumulation plays a great role in the surface discharge under PWM-like excitation.