Short-Term Forecasting of Uncertain Parameters for Virtual Power Plants

Forecasting uncertain parameters are of great importance in the control scenarios of virtual power plants (VPPs). Short-term forecasting of energy consumption, air temperature, and market price is of great importance in the optimization engines in VPPs. Though with the advent of telemetering devices and climate forecasting systems, the energy consumption profiles of consumers and the climate profiles have been collected easily, and consequently the required input data for forecasting these parameters have been available easily, the accuracy of forecasting algorithms has remained an issue. Also, due to the impact of several actors on the market value, no fixed model can be used for all types of markets. This paper focuses on the effect of the time-series analysis methods on developing effective short-term forecasting models for these uncertain variables in a VPP. For this purpose, several models of energy forecasting were built and tested on the Formentera dataset. Our results show that by building the models based on the analysis results, the simple forecaster's models with lower complexity can accurately forecast the uncertain variables of VPPs.