Development of a model for predicting the load on a steam boiler of a thermal power plant

Abstract

As a result, a model for predicting the load at a thermalpower plant was built. To achieve this goal, data on theoperation of the turbine for three years after the plant'soperation were collected. Then, the data were analyzed usingemission processing, trend and seasonality identification, andmodeling of new features. As a result of the analysis, themost correlated time intervals were identified, which wereultimately fed into the model for training. The developedmodel demonstrates high performance in predictionverification metrics.