Identification and extraction of PD current pulse shapes in a cable monitoring system

The most important indicator of the health of electrical plant items is the condition of their insulation for which partial discharge (PD) test or monitoring has been widely recognized as one of the best tools. The strategic importance of plant in the transmission network, and its high replacement and refurbishment cost, has categorized research and development aimed at PD based condition monitoring at transmission voltages (132 kV to 400 kV). This paper presents the signal analysis- a key part for effective condition monitoring, carried out by the present authors on a cable on-line PD measurement system where PDs from the cable under analysis, PDs from local substation plant items and various forms of noise can all exist simultaneously. In this situation, the most critical step lies in the identification and extraction of the pulses associated with cable PD activities. The paper presents an on-line cable condition monitoring system which is installed in a UK utility and the resultant typical data samples. The current PD pulse shapes are categorised into several types using wavelet transform. Based on their characteristics, a method is proposed to separate pulse signals with different shapes. Results presented in this paper prove that the proposed classification method may constitute a step forward in the automatic PD diagnosis.