Estimation of voltage sags from a limited set of monitors in power systems

This paper proposes a probabilistic method aimed at estimating the voltage sag frequency of non-monitored sites given the number of sags recorded by a limited set of voltage sag monitors. The proposed method is based on simulating a representative database of feasible pseudo-measurements by using probabilistic values of fault coefficient, fault impedance, fault type, and fault location. This information is processed through Bayesian filtering based on the Bayes' probabilistic theorem. This allows obtaining the most probable voltage sag index at non-monitored busbars given the index measured at monitored busbars. This statistical estimation of sag indices at locations where there are no monitors constitutes the main contribution of this work. This study is developed in three power systems, the IEEE 24-busbar reliability test system, the IEEE 18-busbar test system and the transmission network of Ecuador with 357 busbars.