Jointly Organized By: Green Energy Management and Smart Grid Research Center (GEMS), Department of Electrical & Computer Engineering, National University of Singapore and IEEE Power & Energy Society, Singapore Chapter and IEEE PES National University of Singapore Student Branch

TOPIC: Solar forecasting: Gaps between Research and Industry

SPEAKER: Dr. Dazhi Yang, Research Scientist: Planning and Operations Management, A*Star (Agency for Science, Technology and Research)

DATE: 22 November 2018, Thursday

TIME: 2.00pm to 3.30pm

VENUE: E4-04-03, Engineering Block E4, Faculty of Engineering, NUS


Despite the significant progress made in solar forecasting over the last decade, most of the proposed models cannot be readily used by independent system operators (ISOs). This is due to three main reasons: (1) the ISOs require the forecasts to be submitted in multi-step-ahead intervals (e.g., thirteen 5-min forecasts over 1 h), whereas the solar forecasting literature focuses on 1-step-ahead forecasting; (2) the ISOs require a lead time for forecast submission, whereas the algorithms used in research almost always ignore such lead time; and (3) the forecast update rate required by the ISOs is generally different from the forecast horizon, the repeated forecasts made to the same timestamps are rarely studied. To that end, this paper proposes an operational solar forecasting algorithm that is closely aligned with the real-time market (RTM) forecasting requirements of an ISO, namely, California ISO (CAISO).

The algorithm first uses the North American Mesoscale (NAM) forecast system, a numerical weather prediction model, to generate hourly forecasts for a 5-h period that are issued 12 h before the actual operating hour, satisfying the lead-time requirement. Subsequently, the world’s fastest similarity search algorithm, a computer-science innovation for time series data mining, is used to downscale the hourly forecasts generated by NAM to a 15-min resolution, satisfying the forecast-resolution requirement. The 5-h-ahead forecasts are repeated every hour, following the actual rolling update rate of CAISO. Both deterministic and probabilistic forecasts generated using the proposed algorithm are empirically evaluated over a period of 2 years.

Dazhi Yang works as a Research Scientist in the research group: Planning and Operations Management in A* Star (Agency for Science, Technology and Research. He obtained his Ph.D. degree from National University of Singapore in 2015. He is now the youngest associate editor of the journal Solar Energy, and is a reviewer for all major energy journals. His research interests include solar forecasting, solar resource assessment, data science, and computer science. In the past five years, he has authored more than 60 papers, for which more than half are first-author journal papers. Dazhi supports open research, to that end, all his first-author papers in the recent years come with data and code. In 2018, he started the Data Article initiative in the Solar Energy journal, first time in 60 years of the journal’s history, promoting world-wide data sharing among solar researchers.