Contents 1 Introduction
Online partial discharge monitoring is a valid means to guard transformer against breakout accident and is very important to power system stably running. In recent years, many scholars and engineers have done larges studies on it and some monitoring devices have been put into test running, however, all monitor devices cannot meet the demand of the insulation diagnosis [1]. At present, the research interests mainly focus on detecting PD ultra-high frequency signal. Larges research work have been done for UHF signal's transmission pattern, spectrum characteristic and collecting method [2]–[5]. The research results show that the UHF method has advantages of anti-interference because its detecting frequency is high enough to avoid most of powerful electromagnetic interferences. It is also proved that UHF signal's spectrum characteristic close correlates to the type of its PD source, which is helpful to PD source identification and location. However, the most of the research is only fit for a single PD source situation. In practice, the situation may be exists that two or more discharge failures occur contemporarily in transformer insulation. The detected UHF signal is probably contributed from two or more PD sources. Hence, it is difficult to realize PD identification and location in the multi-PD situation by scalar sensors.
