Development of computer-based measurements and their application to PD pattern analysis

Statistical Analysis of Partial Discharges from Electrical Trees Grown in a Flexible Epoxy Resin

“…Several statistical parameters were calculated for each distribution, namely mean, standard deviation, skewness and kurtosis. These statistical parameters are considered to describe comprehensively the shape of the corresponding distribution and are widely used in PD pattern recognition [6][7][8][9]. Four different pulse height distributions Hn(q) can be considered taking into account the polarity of partial discharge amplitudes:…”

Conception of complex probabilistic neural network system for classification of partial discharge patterns using multifarious inputs

“…PD is a complex random process (Candela et al, 2000;Gulski, 1995;James & Phung, 1995;Pedersen, Crichton, & McAllister 1991;Van Brunt, 1991), in order to obtain meaningful data for pattern recognition, it is necessary to acquire the fingerprints of the PD signals under well defined conditions for which the cause of PD is known. For an investigation, the PD data is detected by a PD meter (DTM type) which has a frequency band of 1 MHz, with a built in oscilloscope for the display of PD pattern.…”

Artificial identification system for transformer insulation aging

“…Therefore, monitoring the aging status of the transformer will be helpful to prevent system failure (Lin, Wu, & Chung, 2008). Since digital PD detection is developed (Ward, 1992) has extremely high diagnosis accuracy for insulating medium (James & Phung, 1995;Tanaka, 1995), it is an efficient way to well identify the aging status of cast-resin transformer for maintenance or replacement to further prevent insulation breakdown (Farag et al, 1999;James & Phung, 1995;Karthikeyan, Gopal, & Venkatesh, 2008;Tanaka, 1995;Yue, Chen, Cheng, Song, & Xie, 2006).…”