Title :
Analysis of PD patterns with integrated distribution parameters using modular neural networks
Author :
Chang, C. ; Su, Q.
Author_Institution :
Dept. of Electr. & Comput. Syst. Eng., Monash Univ., Clayton, Vic., Australia
Abstract :
Computer based partial discharge (PD) source identification using phase resolved (PRPD) and/or pulse height resolved (PHPD) PD distribution variables is regarded as an important step for insulation diagnosis. Based on those defined in the IEC 270 international standard, distribution variables, derived from both PRPD and PHPD are investigated with a modular neural network approach for PD pattern classification purpose. In this paper, it has been illustrated that with a new PD measuring and analysis system, statistical parameters of PRPD and PHPD distribution variables are practically useful for the identification of PD sources of different type. The applicability of this approach has been demonstrated by obtaining the results from experiments
Keywords :
IEC standards; insulation testing; neural nets; partial discharge measurement; pattern classification; power engineering computing; statistical analysis; IEC 270 international standard; PD patterns analysis; computer based partial discharge; distribution variables; insulation diagnosis; integrated distribution parameters; modular neural networks; pulse height resolved; pulse height resolved PD distribution variables; source identification; statistical analysis; statistical parameters; Computer networks; Detectors; Distributed computing; Frequency; IEC standards; Insulation; Neural networks; Partial discharge measurement; Partial discharges; Pattern analysis;
Conference_Titel :
Power System Technology, 2000. Proceedings. PowerCon 2000. International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-6338-8
DOI :
10.1109/ICPST.2000.900066
