Developing a Comprehensive, System-Wide Forecasting to Support High-Penetration Solar

Developing improved solar irradiance forecasting methods and new and more accurate methods for estimating day-ahead behind the meter solar generation.

Grid Decarbonization and Decentralization Variable Renewable Generation

Clean Power Research, L.L.C.

Recipient

Napa, CA

Recipient Location

3rd

Senate District

4th

Assembly District

beenhere

$744,661

Amount Spent

closed

Completed

Project Status

Download Project

Project Result

The project developed several forecast improvements by advancing methods for identifying low-level clouds motion and PV forecast uncertainty. The researchers refined the Reconstituted Load Model, developed under a prior EPIC agreement, to account for systematic day-ahead forecast errors due to the misspecification of the solar PV generation impact. The study demonstrated that the accuracy of the existing Reconstituted Load models can be improved by replacing the unadjusted solar PV generation with statistically-adjusted data. The day-ahead forecast of the morning, midday/afternoon, evening, and dawn hours are expected to have accuracy gains of 20%, 3%, 10%, and 8%, respectively. CAISO has incorporated this project's BTM forecast into its short-term load forecast models to improve the forecasts as penetrations of BTM solar increase. The final report was published in September 2020.

View Final Report

The Issue

California experiences a wide range of meteorological phenomena, including coastal and valley fog, monsoon events, and temperature inversion and smog events. These phenomena impact the reliability of both satellite-based and Numerical Weather Prediction (NWP) forecast models due to the challenge of predicting cloud formation and dissipation during these conditions. This weather-based uncertainty translates into PV simulation uncertainty in which electric grid operators increasingly rely upon forecast of PV production in their dispatch of operating resources. Forecast inaccuracies cost California millions of dollars annually and result in the unnecessary curtailment of renewable generation.

Project Innovation

This project developed, tested, and validated a high-accuracy forecast for photovoltaic (PV) generation across California and coordinated with the California Independent System Operator (CAISO) on incorporating the results into its PV forecasting operation. The comprehensive forecast included both behind-the-meter (BTM) and in-front-of-the meter scale PV systems. The project quantified the costs and benefits of these improvements. The researchers used mid-term distributed energy resource (DER) adoption forecasts adapted from the investor-owned utility distribution resource plans to project the distribution of DERs through 2050. The team also provided CAISO with the steps required to incorporate the statistically-adjusted BTM solar PV generation estimates into CAISO's Reconstituted Load forecasting approach.

Project Goals

The goal was to improve upon existing solar forecasting methods and develop new ones with greater accuracy.

Project Benefits

This project advanced the state of PV forecasting in California by improving the accuracy of solar irradiance and PV forecasts, particularly rooftop solar, which was not previously well-understood. Participants in the Energy Commission's January 2017 forecasting workshop identified the lack of visibility into DER impacts on net load as a major barrier to generating accurate forecasts. Forecast inaccuracies cost California millions of dollars annually and result in the unnecessary curtailment of renewable generation.

Affordability

This project implemented and improved PV production forecasts that will enable CAISO to reduce net load forecast uncertainty, resulting in cost savings to California.

Environmental Sustainability

Improved solar forecasts improve California’s environmental footprint, because greater DER generation will decrease the use of fossil-fueled reserves, which would otherwise be required to accommodate PV forecast inaccuracy.

Reliability

This project provided CAISO with an improved PV production forecast to increase electric power system reliability across California. The uncertainty of PV generation imparts costs to the California ratepayer.