Short-Term Forecasting Model for Solar PV Power Output using LS-SVM

  • Russel John Gallano Electrical and Electronics Engineering Institute, College of Engineering, University of the Philippines, Diliman, Quezon City, 1101, Philippines

Abstract

The use of renewable energy resources is becoming more prevalent nowadays, especially in distribution systems and microgrids. However, the variability of renewable energy output poses a challenge on the stability and resilience of the power system, particularly in balancing the supply with the load. An output forecast model is useful in this balancing, esp. in scheduling the supply power.

 

Solar photovoltaic (PV) systems, commonly used as a distributed generator (DG), has a variable output that depends on external factors, such as temperature, irradiance, cloud cover, and so on. The lack of data about these external factors may hinder the accurate modelling and forecasting of solar PV output. This study attempts to develop a short-term forecast model of the output power of solar PV DGs using only historical solar PV output data. Least-Squares Support Vector Machine (LS-SVM) is used to establish the forecasting model and shows promising accuracy, even when used to forecast fluctuations in solar output.

 

Keywords forecasting model, LS-SVM, short term forecast, solar power model, solar PV

Published
2021-12-14
Section
Articles