Frequency Selective Multi-Sensor System for Partial Discharge Detec-tion on Medium-Voltage Cable Connectors

Inhalt:
Noise and external interferences in high-voltage (HV) and medium-voltage (MV) environments can lead to wrong partial discharge (PD) diagnosis. The noise signals can be in the same frequency range as the PD signals. Therefore, the identification of the PDs in a harsh environment can be complicated. Furthermore, using IEC 60270 compatible test methods or high-frequency current transformers (HFCT), the localization of the faulty components is often time-consuming and expensive. The novel idea of this work is the use of a frequency selective multi-sensor system, which enables a reliable and cost-effective PD detection on medium-voltage cable connectors. With the combination of an HFCT-sensor and a low-cost capacitive or ohmic sensor, the PD signals of the cable connector can be detected in different frequency ranges. The frequency selective approach increases the reliability of the measurements in a noisy environment. The efficiency of this measurement method in the presence of different noise sources is discussed. In this regard, row measurements of different 10 kV cable connectors with PD faults are performed in an MV switchgear. It is shown that external noise signals are not influencing all sensors in the same way. Furthermore, the comparison of signal amplitudes of the different sensors enables the detection of the local cable connector faults. Based on the proposed analysis method, the PD fault could be classified as local or external to the cable connector even with significant background disturbances.