Statistical Challenges in Solar Power Forecasting

• Activity Number: 414
• Type: Topic Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): YoungDeok Hwang*
• Companies: IBM T.J. Watson Research Center

Abstract

Building a large scale forecasting model for solar energy needs to address some unique challenges. Development of a new operational forecasting system requires understanding the needs of a energy suppliers, as well as practical issues associated with a large uncertainty from the numerical weather prediction model and underlying complex weather system. Also, the model should be able to leverage the presence of multiple physics weather models for forecasting purposes. Not only the accuracy but also a stability of the forecasting system is necessary for large scale operation. Statistics has an important role in quantifying uncertainty and providing guidance to decision-makers in the market. I will discuss some of these unique challenges and research opportunities in applying statistical methods in solar energy industry.