Projects
Quickly jump to Turbulence dissipation rates | Mesoscale Modeling of Wind Farms | Flows in Complex Terrain | Marine Boundary Layers | Observations of Wind Turbine Wakes | Atmospheric Impacts on Energy Production | Large-Eddy Simulations (LES) of the Atmospheric Boundary Layer (ABL) | Wildfire Modeling | LES of the Hurricane Boundary Layer | Mesoscale-microscale Coupling | Assessments of boundary-layer instrumentation | Climate Change Impacts on Energy Production ´¥ÌýUrban Meteorology
A fundamental challenge in boundary-layer meteorology is how turbulence is dissipated; inaccuracies in the balance of turbulence production and dissipation undermine our ability to simulate atmospheric flows where turbulence is important. With support from NSF via Lundquist's CAREER award, weÌýhave developed and demonstrated techniques using lidar to estimate dissipation rate in flat terrain (), in complex terrain (), in complex terrain with wind turbine wakes () and offshore (). We used complex terrain datasets to test a machine learning approach for approximating dissipation rate (). We thenÌýexpanded this approach for a different complex terrain dataset and other lidar instruments (). To better represent turbulence dissipation in numerical models, we are testing existing models () and new modeling approaches (, ).
Ìý
Mesoscale Modeling of Wind FarmsÌý
To assess the local and regional impacts of wind energy development, we have implemented a wind farm parameterization into theÌý. The Fitch et a. parameterizationÌý) is available with every WRF download since version 3.3. In simulations of the CASES-99 GABLS case, the wind farm wake varies throughout the diurnal cycle, with the maximum downwind surface temperature increases ~ 0.5K at nightÌý) consistent with observations. Comparisons between this elevated drag model and climate simulations representing wind farms simply with enhanced surface roughness show nearly the opposite local impacts on surface temperature (). Even in the midst of a very large wind farm, the type of crop surrounding the turbines can impact the wind resource (). Comparisons of these simulations with large-eddy simulations suggest that inclusion of turbine-generated turbulence is essential (). Using wind farm power production data, we have shown that the use of the wind farm parameterization improvesÌýforecasts of power production (). Large-scale climate impacts of wind energy deployment are still being explored (). We have extended our wake analysis work to validate modeling tools by comparison to field data (, , , , and ), tested a rotor-equivalent wind speed formulation in (), provided the scientific community guidance on best approaches for simulating wind plant wakes (), demonstrated how wakes can affect weather like thunderstorms (), quantified the financial impact of wakes on neighboring wind farms (), and demonstrated the effects of wakes on long-term observational capabilities ().Ìý
The atmospheric boundary layer behaves very differently in complex terrain, with impacts on wind energy and phenomena like mountain venting. We have recently participated in three experiments in complex terrain. TheÌýDOE-funded WFIP2 project sought to improve forecasting models in complex terrain (, , and ). One novel aspect of the WFIP2 observations was the documentation of how mountain waves affect wind energy production (, ). We have used these data to test and extend dissipation rate measurements in complex terrain (). With collaborators, we are using these data for assessing novel methods for modeling flows in complex terrain ( including statistical forecasting methods () and operational models (Pichugina et al. 2022 in press WAF, , , , ). In the international wind energy complex terrain ±Ê±ð°ù»å¾±²µÃ£´Ç experiment (Portugal 2017, ), we deployedÌýa tethered lifting system for unique in-situ measurements of turbulence (TLS) as well as ground-based lidars. We used lidars to characterize complex flows such as recirculation () and compared lidar and TLS measurements of dissipation rate (). Using the TLS and lidar measurements, we evaluated mesoscale-microscale simulation approaches () in complex terrain. Finally,Ìýa small NREL/industry complex-terrain experiment provided new insights for testing wind turbine control strategies in complex terrain, summarized in , . We used this dataset () to demonstrate the how the atmospheric profile can dictate wind turbine power performance, in contrast to work we have done in simple terrain ().
Beyond extreme events like hurricanes, we are also assessing the characteristics of the marine boundary layer relevant for wind energy. We assessed offshore wind shear and veer profiles (). and turbulence dissipation rate data (). We quantified how many offshore wind turbines would be required to match electrical demand in the US Northeast (), and assessed which reanalysis datasets are best suited for modeling the offshore environment (). We are funded to collect and analyze a broader dataset via an upcoming Dept. of Energy experiment (WFIP3) for which my students and I will deploy our instrumentation and carry out numerical weather prediction simulations.ÌýWe have extended our work in wind farm wakes to the offshore environment via comparisons to European offshore wake measurements (, , and ). We have also simulated wakes from the planned US offshore wind plant sites, including uncertainty quantification (Rosencrans et al. 2022, in preparation) and multiple boundary-layer models ().
Observations ofÌýWind Turbine WakesÌý
We utilize profiling lidar (,Ìý,Ìý, ,Ìý), scanning lidar (,Ìý,Ìý,Ìý, , Ìý), radiometers (, ), meteorological towersÌý,Ìý) and tethered lifting systems () to study the development and propagation of wind turbine wakes in different atmospheric conditions, with special interest in wind farms co-located with agriculture (,Ìý). Our algorithms for wake characterization are publicly-available at .
Ìý
Ìý
Atmospheric Impacts on Wind Energy ProductionÌý
Atmospheric stability impacts wind turbine power production (, ) sometimes in contradictory ways (ÌývsÌý) depending on local meteorology. Neural networks can extend limited measurements at a site towards improved resource assessmentÌý). Detailed lidar observations and WRF simulations of nocturnal low-level-jets (LLJ) () suggest that LLJ-induced wind shear and veer extend into the turbine rotor-layer during intense jets. This wind shear and veer varies with stability and impacts power production (, )ÌýAt large spatial scales, we evaluate the statistical independence of wind generators, and find that higher-rate fluctuations in wind power generation can be effectively smoothed by aggregating wind plants over areas smaller than otherwise estimated (). ForÌýwind resource assessment of P50 and P90 levels, interannual variabilityÌýis influenced by data record length, with counter-intuitive suggestions for ideal record length (). We assess methodologies for long-term wind resource assessment (). Further, wind turbine nacelle measurements are also affected by atmospheric stability () as is power production ().
Improved turbulence modelsÌýb) and/or immersed boundary methodsÌý,Ìý) may be required to simulate complex flow in stable atmospheric boundary layers or in regions of complex terrain. To elucidate interactions of atmospheric stability with wind turbine wakes, large-eddy simulations with a generalized actuator disk model (,Ìý, ) demonstrate how ambient turbulence accelerates the erosion of wind turbine wakes. These LES of turbine wakes can be used to evaluate mesoscale parameterizations of wind farm effects ().ÌýAdditionally, we validated LES in comparison to field data for dispersion experiments ()Ìýand for heterogeneous experiments incorporating wind turbines in the flow (, Sanchez GomezÌýet al. 2022 in review JRSE).ÌýWe use finer-resolution simulation tools to represent individual turbines and can assess the impact of the wind profile and atmospheric stability on wakes () and how those wakes decay and affect downwind environments (). The wind veer and wind shear characteristic of nighttime environments interacts with the direction of turbine rotation in complex ways (; ).
Smoke lofting from a fire changes with varying local winds, relative humidity, and atmospheric boundary-layer stability, and the presence of moisture can raiseÌýthe altitude of lofting, while faster wind speeds dampen loftingÌý().
Ìý
Ìý
To better understand the environment for offshore wind turbines, detailed large-eddy simulations can provide information about turbulence structures that observations simply cannot. We have developed () and verifiedÌý() a simpler LES approach and used these simulations to quantify gust factors for wind turbine design () to help improve turbine design standards for offshore wind turbines. Working with structural engineers, we have used these data to assess loads on wind turbines ().Ìý
Ìý
Ìý
Simulations provide an important tool for understanding and predicting the atmospheric boundary layer, but we need to incorporate both realistic large-scale weather effects (mesoscale) as well as refined localized turbulent variability (microscale) with large-eddy simulations, transcending previous idealized approaches. We have investigated and tested new approaches for simulations across scales (, ). We have demonstrated such couplingÌýfor a diurnal cycle in a wind farm, facilitated with cell perturbation methods ().
Ìý
The US Dept of Energy supported an experiment, XPIA, to assess the ability of current instrumentation to measure wind farm flows (). This dataset allowed for detailed investigations of the performance of radiometers ( ), scanning lidars (,Ìý) and their uncertainty (), comparisons of scanning lidars with dual-Doppler radar (), and even meteorological tower wakes (). Data from the XPIA campaign may be downloaded fromÌý. We have assessed lidar capabilities for measuring wind turbine wakes (), quantified measurement strategies for airborne lidar measurements (), and developed a lidar simulator () and lidar error assessment tool that is now in use for designing upcoming field experiments (Sanchez GomezÌýet al., 2022, in review at JRSE).
Climate Change Impacts on Energy Production
As the climate changes, our understanding of how to assess long-term renewable resources () and future predictions of those resourcesÌý() will change. As a result, weÌýneed for incorporating weather and climate data into energy systems modeling (, ).
Cities can create their own microclimates with interesting implications for detailed simulations of flow in urban areas (;Ìý), mixing-length heights (), and air pollution events (), especially when urban meteorology interacts with nocturnal low-level jets (LLJ). Immersed boundary methods may be required for simulations in urban areas (, ).