Risk modeling of accidents in the power system of Ukraine with using bayesian network

Abstract

Current studies of impact of climatic factors on overhead power lines are limited to calculations of load of climatic factors on the overhead transmission lines, so the problem of conducting a comprehensive study of accidents probability under the influence of climatic factors is important. The paper addresses the research of approaches to spatial risk modeling of overhead power lines accidents in the power systems of Ukraine under the influence of climatic factors. The article presents the construction of a model of accidents under the influence of climatic impacts and prediction of emergencies on based geospatial data sets. Pattern recognition techniques, namely the Bayesian network, were used to simulate accidents and verification of the results. This method is based on calculation of a posteriori probabilities of model variables. As a result, a model of accidents under the influence of climatic factors was built, which constitutes a Bayesian network with given conditional probabilities and independent variables of the model.