Using machine learning methods to forecast the number of power outages at substations

Forecasting in the energy sector is of great importance for suppliers and consumers. Optimum power consumption depends on many factors. Due to natural or any other external conditions, accidents are possible. In order minimize emergency consequences, it necessary be prepared possible outages advance reduce time their elimination decision-making. This article considers problem forecasting at substations. The enterprise provided a summary table substations due disasters specific days. To solve problem, machine learning method was chosen – binary classification. Five different algorithms were considered. models tested data from first half 2022. most effective algorithm 20% test sample classification using generalized additive (GAM). also one best with 50%. A model has been further use predicting probability enterprise. can used organizations; this, train corresponding region.