| Analyze generator stator insulation inner discharge mechanics, establish the ideal mathematical models of the three type discharges, and design the samples who will be used of generator partial discharge on-line pattern recognition ;Adopting layers anti-jamming method process the data who are collected by generator partial discharge on-line monitoring system; Obtain generator partial discharge statistical characteristic vectors and polarity statistical atlas; Design generator discharge on-line pattern recognition separate based on the polarity statistical atlas and the statistical vectors and establish the neural network classifier who will be used of on-line pattern recognition; Exploit generator partial discharge on-line pattern recognition system. |
