PV power forecasting using a parametric model

Abstract

Nowadays, accurate AC power output forecast of PV plants can play a significant role for both the system operators and PV plant owners. The first group has to program the dispatch of the electric system and the second one needs to plan their energy commercialization schedule up front. Therefore, after analyzing the existing commercial services of PV production forecast, it has been identified the potential to improve the models used to calculate the AC power output from irradiance and ambient temperature forecasts. Then, in this paper, a parametric model is proposed in order to improve the forecast performance. This model provides the AC power at the output of the PV plan using as input variables predicted ambient temperature and predicted global horizontal irradiance, both obtained from open sources. It has been evaluated and validated using historical production data of a fixed generator (2,24kWp) and a large PV power plant (45MWp) and historical forecast data from the open source Meteogalicia In addition, a toolbox to implement this parametric methodology has been built upon the Matlab environment. The toolbox has been developed to provide the PV plant production one day in advance.