Integrating large amounts of solar power into the existing electric grid is problematic because power is only produced during daylight hours, and can be interrupted by passing clouds. This intermittency leads to a mismatch between power production and power consumption, which limits the adoption of solar power by utility companies and large-scale power users who require a stable and reliable power source. Discovering new ways to compensate for intermittency will accelerate the adoption of solar power, will reduce CO2 emissions, and will improve national security by conserving non-renewable resources.
To accommodate intermittency, we will regulate industrial-scale electrical demands on a minute-by-minute basis in response to measured and forecasted solar power. We will conduct experiments at the Biosphere 2 in Arizona, where we will use measured and predicted output of 70 kW of solar PV modules present at the site to regulate the operation of two 75 HP (56kW) water pumps. The water pumps transfer water from two wells into a 500,000 gallon reservoir. If the activity the water pumps is suitably regulated based on the power produced by the PV modules, the variability in the power production of the combined system (PV + Pumps) will be reduced compared to that of the PV modules alone. A great deal of theoretical work has been done in this field; the proposed project will provide much needed experimental data.
Broader Impacts
Findings from our experiments will inform large power users such as the University of Arizona, the city utility Tucson Water, and the Central Arizona Project (CAP) about cost-effective and safe ways to use more solar power. It will also enable industries to plan bigger solar farms without net-metering. The graduate and undergraduate students will benefit from the participation.
Integrating large amounts of solar power into our electricity grid is challenging because of the variable nature of this renewable resource. Forecasts of the output from solar power plants will help with this challenge, for example by enabling industries to adjust their loads to better match the schedule of available solar power. Forecasts can also help inform utilities so they can maintain appropriate amounts of backup power. To address this challenge, we developed improved forecasts for solar power plants by a variety of methods. We pioneered short term (intra-hour) forecasts of solar power using data from a network of irradiance sensors. We demonstrated a proof of principle by using historical data from 100 rooftop systems. We presented this result in 2012 [1]. To support our analysis, we developed methods to identify site-specific clear-sky expectations for power output [2]. More recently, we published this method in a journal article [3] and developed new hardware and communications to forecast the output from utility scale PV power plants in real-time [4]. We analyzed the variability of several solar power plant output in Arizona. We studied the size of batteries to provide ramp rate control for large or small PV power plants [5]. We studied how curtailment schedules informed by our solar power forecasts can limit ramp rates [6]. We also used historical data from 82 MW of solar power plants in Tucson and 500 MW of solar power plants in Arizona [7] to calculate recommended schedules of spinning reserves to compensate for fluctuations in the output from these geographically distributed portfolios of PV power plants [8]. Numerical weather models are another tool for renewables forecasting. We use outputs from a WRF model to forecast PV power and wind power generated by Tucson Electic Power assets. A description of this system was presented in [9] and [10]. Intellectual Merrit: We compared the RMS, MBE and MAE errors of ours and other forecasting techniques. We reported the time range (forecasting horizons) for which our methods perform better than the persistence model. We investigated the origin of errors, such as incorrect assumptions for wind velocity, differences between cloud velocity and wind velocity, and occasional situations with clouds at multiple altitudes. We are now studying how small the errors can be at various forecasting horizons. We also investigated how to use forecasts for ramp rate control and scheduling of spinning reserves. Broader Impact: Students and postdocs trained during this research have learned to work with solar power installers and utility companies. Advanced solar power forecasts will help our utilities incorporate larger amounts of renewable energy, thus saving fossil fuels and emitting less carbon dioxide, while maintaining a reliable electricity grid. [1] V.P. Lonij, V.T. Jayadevan, A.E. Brooks, K. Koch, M. Leuthold, and A.D. Cronin, "Improving forecasts of PV power output using real-time measurements of PV output of 100 residential PV installs," Proceedings of IEEE Photovoltaics Specialists Conference (2012) [2] V.P.Lonij, A.E.Brooks, K. Koch, and A.D. Cronin, "Analysis of 100 rooftop PV systems in the Tucson, AZ area," Proceedings of IEEE Photovoltaics Specialists Conference (2012) [3] V.P.A. Lonij, A.E. Brooks, A.D. Cronin, M. Leuthold, K. Koch, "Intra-hour forecasts of solar power production using measurements from a network of irradiance sensors" Solar Energy 97, 58 (2013) [4] Antonio T. Lorenzo, William F. Holmgren, Michael Leuthold, Chang Ki Kim, Alexander D. Cronin, Eric A. Betterton, "Short-Term PV Power Forecasts Based on a Real-Time Irradiance Monitoring Network" Proceedings of IEEE Photovoltaics Specialists Conference (2014) [5] Daniel Cormode, A.D. Cronin, William Richardson, Adria Brooks, Daniella DellaGiustina, Antonio Lorenzo "Comparing ramp rates from large and small PV systems, and selection of batteries for ramp rate control", Proceedings of IEEE Photovoltaics Specialists Conference (2013) [6] Daniel Cormode, Antonio Lorenzo, Will Holmgren, Sophia Chen, and Alex Cronin, "The Economic Value of Forecasts for Optimal Curtailment Strategies to Comply with Ramp Rate Rules" Proceedings of IEEE Photovoltaics Specialists Conference (2014) [8] http://sveri.uaren.org is an interactive website we developed for the Southwest Variable Resource Initiative (SVERI) to show data from variable energy resources operated by several utilities in the desert southwest U.S. This site permits downloading data for a user-controlled range of dates. [7] Daniel Cormode, Antonio Lorenzo, Will Holmgren, and Alex Cronin, "Observed Fluctuations in Output from a Regional Fleet of PV Power Plants used to compute hourly schedules of Spinning reserve requirements" In preparation for publication in the EU PVSEC conference, September (2014) [9] William F. Holmgren, Antonio T. Lorenzo, Michael Leuthold, Chang Ki Kim, Alexander D. Cronin, and Eric A. Betterton, "An Operational, Real-Time Forecasting System for 250 MW of PV Power Using NWP, Satellite, and DG Production data" Proceedings of IEEE Photovoltaics Specialists Conference (2014) [10] Solar Electric Power Association (SEPA) Report by Seteven Letendre, et al. "Predicting Solar Power Production: Irradiance Forecasting Models, Applications, and Future Prospects", March (2014)
