Machine learning for solar irradiance forecasting of photovoltaic system Jiaming Li a, * , John K. Ward b , Jingnan Tong b , Lyle Collins b , Glenn Platt b a CSIRO CCI, Australia b CSIRO Energy Technology, Australia article info Article history: Received 7 September 2015 Received in revised form 24 November 2015 Accepted 31 December 2015 Available online 18 January 2016 Keywords: Virtual power station Renewable energy optimization Genetic algorithm Hidden Markov model SVM regression Neural networks abstract Photovoltaic generation of electricity is an important renewable energy source, and large numbers of relatively small photovoltaic systems are proliferating around the world. Today it is widely acknowledged by power producers, utility companies and independent system operators that it is only through advanced forecasting, communications and control that these distributed resources can collectively provide a firm, dispatchable generation capacity to the electricity market. One of the challenges of realizing such a goal is the precise forecasting of the output of individual photovoltaic systems, which is affected by a lot of factors. This paper introduces our short-term solar irradiance forecasting algorithms based on machine learning methodologies, Hidden Markov Model and SVM regression. A series of experimental evaluations are presented to analyze the relative performance of the techniques in order to show the importance of these methodologies. The Matlab interface, the Weather Forecasting Platform, has been used for these evaluations. The experiments are performed using the dataset generated by Australian Bureau of Meteorology. The experimental results show that our machine learning based forecasting algorithms can precisely predict future 5e30 min solar irradiance under different weather conditions. © 2016 Elsevier Ltd. All rights reserved. 1. Introduction The world has abundant solar energy resources. Photovoltaic (PV) technology has become one of several promising alternatives for use in energy technology [1]. Yet many critics of the widespread use of solar energy cite its intermittency, or the challenges around predicting the future output of a solar generator. The Virtual Power Station (VPS) [2e4] conducted by CSIRO aims to address such concerns by combining a large number of geographically disperse, and technically diverse, small scale renewable energy generators that will allow them to present to the electricity market as a single reliable dispatchable entity. The aggregated energy of the VPS can be sourced from a large number of small energy generation and storage systems, such as roof-mounted solar PV panels, and asso- ciated grid-connected battery systems installed in individual do- mestic houses. These individual systems are then aggregated together, to form a “virtual power station”, with one coordinated response, of benefit to the wider electricity network. However, integration of large amounts of PV into the electricity grid poses technical challenges due to the fluctuating characteristics of avail- able solar energy sources. PV output is not easily predictable in advance and varies based on both weather conditions and site- specific conditions. Such variability of solar energy resources at ground level thus raises concerns regarding how to manage and integrate output from the VPS to the power grid. Given the issues above, there is increasing interest in more precise modeling and forecasting of solar power. Irradiance is a measurement of solar power and usually measures the power per unit area. Most researches consider the solar irradiance forecasting at a site, which is essentially the same problem as forecasting solar power. The ability to forecast solar irradiation will enable power grid operators to be able to ensure the quality and control of solar electricity supplies in an environment of greater solar panel usage, allow them to better accommodate highly variable electricity generation in their scheduling, dispatching, and regulation of po- wer. In particular, the possibility to forecast solar irradiance can became fundamental in making power dispatch plans, and also a useful reference for improving the control algorithms of battery charge controllers. Ultimately, the development of more accurate * Corresponding author. E-mail address: jiaming.li@csiro.au (J. Li). Contents lists available at ScienceDirect Renewable Energy journal homepage: www.elsevier.com/locate/renene http://dx.doi.org/10.1016/j.renene.2015.12.069 0960-1481/© 2016 Elsevier Ltd. All rights reserved. Renewable Energy 90 (2016) 542e553