Grid-integration of distributed generation sources continues to be a hot topic for utilities and independent system operators (ISOs). Last month, the Utility Variable Generation Integration Group (UVIG) held its 6th annual workshop on Variable Generation Forecasting Applications to Power System Planning and Operations. At this event, numerous wind and solar energy forecasting topics related to power system operations were discussed from both U.S. and European perspectives.
Several excellent talks by European utility and forecasting entities focused on regional-specific concerns, foreshadowing challenges the U.S. is likely to face as PV penetration increases. According to Dr. Alejandro J. Gesino from Amprion GmbH, Germany currently has approximately 31 GW of both wind and solar generation resources installed (approximately 62 GW total), and system integration is continuing to challenge Regional Transmission Organizations (RTOs) on a daily basis.
While most German wind energy is utility scale and tends to be metered, the vast majority of solar energy in Germany is behind-the-meter in small distributed generation (DG) installations. Unlike metered PV systems, utilities have no direct way to monitor the energy production of systems installed behind-the-meter. As a variable energy source, deriving and predicting the energy produced by behind-the-meter solar is daunting at best, and will be something of a wild card as the level of installed capacity of DG solar continues to increase in Germany.
Although the penetration of behind-the-meter solar in the U.S. has not yet reached German levels, forecasting solar is becoming a challenge for several Western U.S. utilities and the California Independent System Operator (CAISO). Clean Power Research founder Tom Hoff described our efforts to help utilities and ISOs address the challenges involved in DG solar generation using SolarAnywhere® FleetView® fleet forecasting capabilities.
The same week, the Department of Energy (DOE) held a half-day workshop focused on solar forecasting metrics needed by utilities and ISOs for load balancing and other grid integration activities. The workshop led to a lively discussion between all stakeholders in the process, ranging from forecasting vendors, to utility planners and system operators.
One area of agreement is the need for benchmarks or comparisons of error metrics in various solar forecasts. Last year, Tom Hoff co-authored a paper with representatives from the University of California, San Diego, SUNY-Albany, NREL and Sandia National Labs discussing this hotly debated topic. The paper describes and compares fundamental ways to benchmark various errors of solar energy forecasts.
While it’s clear that forecasting of DG solar will be needed to integrate high-penetration PV, ultimately it will be energy trading market conditions and needs that will determine which solar forecasting metrics will be most useful for each market.