Bayesian Networks for Power Quality Analysis in the Industrial Sector

This article presents a model to evaluate the power quality phenomenon denominated voltage sags by means of Bayesian networks. This model considers the main events that can be generated by the presence of voltage sags in the power industrial systems, especially on electronic and sensitive equipment. The relations of causality established between the main causes and the final effects consider both the internal and the external electrical conditions of the industrial power system, such as environmental effects. The probabilities of the causality relations are assigned from statistical data available in literature, from Monte Carlo simulations, sensitivity analysis and expert opinion. The model is very versatile, since it analyzes the sag phenomenon for any type of industry and voltage level and its impact on possible actions or corrective measures in aspects such as: maintenance, selection and coordination of protections, etc. in order to correct or to reduce the voltage sag phenomena and improve the service quality to the client and to other clients connected to the same distribution circuit