Voltage Sags Duration Probability Distribution Function

This paper aims to present a probability distribution function for voltage sags. A large number of real grid recordings, originally obtained by authors, measurements provided by Prof. Math Bollen or presented in the literature are analyzed. Two types of probability functions, Weibull and lognormal are considered for five cases of measurement results from different countries. Averaged probability distribution functions are proposed. Their applications for the optimization of the artificially generated voltage sags database can be used further for the training of neural networks for power quality characterization.