Voltage sag analysis on three phase systems using wavelet transform and probabilistic neural network

摘要: The ever growing diffusion of susceptible loads in power system result increasing susceptibility to quality problems. For this reason data are recorded by utilities with the aim identifying origins these disturbances order improve reliability electrical and thus reduce their impact on customers. In paper a classification algorithm for automatic analysis event is presented. approach based wavelet probabilistic neural network classificator. discrete transform (DWT) used detect fast changes voltage signals. It possible carefully starting recovery times disturbance other features, such as start slew rate intermediate number phases involved each fault event. These features input patterns second step (PNN) classify sag events into four classes (network faults, motor starting, re-acceleration after fault, transformer energising).