Assesment of the Performance of the Neural Model of the National Electricity Demand System

Abstract

This paper presents the results of research on the use of a neural model’s capabilities for forecasting national power demand using data from the years 2019–2023. The neural model is a daily MIMO-type model with 48 hourly simultaneous inputs concerning the total electricity generation by the JWCD system and by the nJWCD system, and with 24 hourly simultaneous outputs concerning national electrical power demand shifted by one day ahead relative to the input quantities. The neural model of the National Power Demand System as well as the simulation model built on its basis for investigating model accuracy in the scope of forecasting were designed and implemented in the MATLAB and Simulink environment. Many interesting research results were obtained, particularly in the scope of model quality.