New weather forecasting tools could aid grid reliability as solar PV is added
The New York Power Authority (NYPA) and its research partners completed a multi-year study to help the state’s solar industry deploy weather forecasting technology to better anticipate power generation and improve electric grid reliability.
The project addressed challenges raised by the uncertainty related to solar output by offering advanced forecasting methods and making a roadmap to help maintain grid reliability, optimise production of renewables and reduce operating costs.
High quality weather forecasting models “will be vital to the operations of utilities and independent system operators,” said Alan Ettlinger, senior director of the New York Power Authority’s Research, Technology Development and Innovation team.
The study showed how how more extensive data and advanced solar-focused models were able to increase the degree of accuracy and granularity that will be needed to maintain grid reliability and support operations.
The study, the final $1.5 million phase of a $2.4 million project, was funded by NYPA, the New York State Energy Research and Development Authority, and the Department of Energy’s Solar Energy Technologies Office, and co-managed by EPRI, an independent, non-profit energy R&D institute.
Also of interest:
#DISTRIBUTECH: Weatherproofing the grid
Predicting space weather to protect the grid
Other partners included National Center for Atmospheric Research (NCAR), Brookhaven National Lab and the State University of New York at Albany.
High-definition digital cameras were deployed in sky-imager networks and advanced weather forecasting models focused on solar forecasting, combined with advanced data from the NYS Mesonet, a statewide network of weather stations, and other resources. Advanced predictive methods were developed and evaluated against forecasts currently in use to show the benefits of more detailed models and data.
The study relied on an open-source, gridded solar power forecasting system. NYSolarCast makes predictions of global horizontal irradiance (GHI) every 15 minutes for a three-kilometer grid covering all of New York.
Those predictions were then used to predict solar power generation for both utility-scale photovoltaic (PV) plants and distributed (mostly rooftop) PV installations. NYSolarCast used machine learning techniques trained on NCAR-based weather prediction models, NYS Mesonet observations and historical data from PV plants across New York.
The study helped develop an underlying platform for solar and other utility-related weather forecasting, including building load management, based on improved solar irradiance forecasts.
Results indicated that the solar industry needs to take steps to be more transparent by sharing data, having better maintenance and monitoring instrumentation, and filtering out erroneous data.
The new models could also form the basis of improved commercial tools. Several companies currently provide forecasting services for New York State, particularly for day ahead operations based on weather modeling. Improved forecasts could be applied to individual solar plants and to predict distributed solar across a large region, not only for NYISO, but also for generation, transmission and distribution companies, private developers, and end user customers.
The report is available here.
