AWAKEN/Science Goals

From Open Energy Information


Home | Research Questions | Location | Instrumentation | References | Funding


Wake losses and wake model uncertainty are a central issue to the US Department of Energy’s (DOE’s) A2e program. To improve wake model prediction accuracy and reduce uncertainty, new validation data sets are needed across a range of scales including utility scale testing within operational wind plants. To meet this need, DOE plans to fund a large utility scale field campaign – The American WAKe ExperimeNt (AWAKEN) starting in late 2020. The field test in currently envisioned for the US Midwest where the largest percentage of existing wind farms in the US operate.

The science goals for the A2e program have been identified at a high level during previous strategic workshops (Hammond et al. 2015) under a process using Phenomena Identification Ranking Tables (Hills et al. 2015). The major phenomena important for wind plant flow physics are identified below and many have been identified as high priority (Maniaci 2016). To further the planning process for a new test, an AWAKEN planning meeting will be held March 21, 2018 in Boulder, Colorado with a group of worldwide experts. The experts will help refine the critical scientific goals and questions to be answered through a new test based on their experience.

Prioritized science goals from breakout sessions at first AWAKEN meeting.

Science Goals and Detailed Questions

  1. Wake recovery and dissipation
    1. How does the wake recovery rate change with atmospheric stability throughout the wind plant? How does stability within the wind farm evolve over the diurnal and seasonal cycles in contrast to concurrent ambient stability?
    2. How does wake-added and ambient turbulence affect the recovery rate and expansion?
    3. What is the best way to calculate or parameterize and observe wake recovery?
    4. How and why is turbulent dissipation in a wake modified compared to the free stream, and how can it be measured and modeled? How does dissipation rate affect wake recovery?
    5. How does the wake width change with background conditions, especially the stratification and shear?
    6. How does available momentum above wind farm impact wake recovery?
    7. How is wake recovery related to wake meandering?
    8. How does yaw angle influence wake recovery and dissipation under different stratifications?
  2. Wake interaction, merging, meandering
    1. How do wakes merge under different atmospheric conditions?
    2. How is the meandering affected by turbine separation distance?
    3. What are the dominant contributing factors to wake meandering? How does atmospheric conditions contribute to wake meandering? How does turbine design contribute to wake meandering? How Does turbine operation contribute to wake meandering? Can these various effects be separated in a measurement campaign?
    4. How does turbine yaw misalignment influence meander and merger?
    5. How does meandering change between the upstream wakes (for the front row of turbines) and wakes deep in the plant?
    6. Is meandering driven only by the large scales in the flow, or does TKE redistribution within the wake (and therefore its expansion) also contribute to wake motion? What is the influence of topography?
  3. Wake impingement on downstream turbines
    1. How do wakes influence blade, tower, and nacelle loading and acceleration on downstream turbines?
    2. How is wake influence impacted by different wake mixing scenarios (single wake vs. mixed wakes from multiple turbines)? How does that change in partial-wake scenarios?
    3. How do yawed wakes impinge differently than standard wakes?
    4. At what wind speeds and other atmospheric conditions do wakes have the greatest loading impact on downstream turbines?
    5. Is the half wake situation the most detrimental for loads?
    6. What are the important loads for wake interaction? Can loading impacts be best isolated in a wake experiment, in combination with simulation or with simulation alone?
    7. How can the inflow - wind turbine - wake system be modeled effectively to predict loads on downstream turbines?
  4. Deep array effects/internal boundary layer
    1. How would a deep array effect be seen and detected in a land-based wind farm?
    2. What are the dominant physical processes required to model the deep array effect for a land-based wind farm?
    3. How does the deep array effect change with atmospheric stability?
    4. How does the deep array effect change with different turbine spacing and layout?
    5. How does the deep array effect change with turbine operation?
    6. What is the shape of internal boundary layer growth above and on the sides of a wind farm? And how is this related to the ambient boundary-layer depth?
  5. Atmospheric stability and surface heat flux
    1. How do the dynamics of a single wake and of the wind plant (intensity/recovery/meandering) change under clear-sky, radiative forcing-driven atmosphere vs. under specific weather events?
    2. What is the impact of lateral coherence (and its height dependence) in the freestream atmosphere on an unwaked turbine vs. a waked turbine (e.g. in terms of yaw loading) and how does it vary with turbine separation?
  6. Momentum transport within, around, above and below the farm
    1. How does boundary layer height (and the development of a wind farm internal boundary layer) affect momentum entrainment above the farm? How does this depend on stability conditions?
    2. Can the influence of the wind farm on momentum be separated from the atmospheric dynamics? In simulation or observation?
    3. Does the speed reduction of the wind farm lead to surface convergence and enhanced upward motion?
  7. Wind direction, shear and veer
    1. How does the wake behavior change under changing shear and veer?
    2. How does wake move under sudden changes in wind direction? What is a quantitative measure of direction change vs. a turbulent gust? Is IEC a sufficient definition?
    3. Surface roughness
    4. How do seasonal changes in roughness affect power performance and plant aerodynamics in general?
    5. How important is it to accurately model roughness in our wind plant aerodynamics simulations? How dependent is the surface roughness impact to other local features such as terrain? How can individual roughness and terrain impacts be isolated in measurements?
    6. How is LCOE affected by roughness management (e.g., corn vs. soy, dormant season roughness management)
  8. Wind plant wake
    1. How far downstream does a wind plant wake persist? How can this be defined if the background wind field is changing?
    2. How does the intensity/fetch of the plant wake vary with atmospheric conditions/terrain/obstacles/plant operating conditions?
    3. How is the plant wake defined? Are there near and far wakes with different physical descriptions and influences as with individual turbine wakes?
    4. Are their downstream turbine loading impacts from wind plant wakes or is the dominant effect power loss?
    5. How does the wind plant wake affect the local climate downstream?
    6. Does a wind plant wake meander?
    7. How can a wind plant wake be best observed?
  9. Terrain impacts
    1. How does terrain affect wake recovery and trajectory (of a single wake vs. the entire plant)?
    2. How does terrain affect loads? If increased shear/turbulence reflect in increased fatigue loads, do terrain-induced speed-up effects outweigh the decreased fatigue lifetime of components/turbine?
    3. At which resolution does the terrain need to be modeled so that its speed-up effect and its effect on wake dynamics are reproduced?
  10. Wind plant upstream blockage
    1. How far upstream is the wind speed changed by presence of the wind farm?
    2. How is the blockage effect influenced by turbine size, spacing, operation and layout?
    3. How does the wind accelerate around the wind farm?
  11. Air-sea interaction (AWAKEN is envisioned for US Midwest, but if offshore is prioritized highly these science goals can be incorporated)
    1. Relationship between wave height and period on tower loads (and platform/anchor system loads for FOWTs), on the flow within the plant, on wake recovery...
    2. How far off are we when we use similarity theory to estimate surface fluxes (from buoy or satellite or model data) and how does this propagate to our extrapolation to hub heights?
    3. Should we revisit how we model (or not model...) offshore roughness?
    4. What are the consequences of using equivalent neutral winds (satellite product) instead of stability-dependent winds when estimating vertical wind profiles and extrapolating low wind speeds to nominal hub heights? I believe that corrections accounting for stability are based on similarity/models.
    5. How do swell conditions affect the atmospheric surface layer, wake recovery, and the turbine power/loads?
  12. Reducing Uncertainty
    1. How will measurement sites be selected to construct ensembles of multiple experiments (e.g. single wakes, merged wakes, wind plant wakes, wind plant-wind-plant interactions)?
    2. How will the selection of sites help reduce measurement uncertainty?


Hills, R., Manicaci, D., and Naughton (2015) V&V Framework, Sandia National Laboratories, SAND2015-7455.

Hammond, S., Sprague, M., Womble, D. and Barone, M. (2015) A2e High Fidelity Modeling: Strategic Planning Meetings, NREL/TP-2C00-64697. Maniaci, D. (2016) Overview of the V&V Framework, Presentation at the Wakebench, IEA Task 31 Meeting, March 16-17, 2016, Albuquerque, NM, SAND2016-2862PE.

Recommended Reading

Iowa Atmospheric Observatory (IAO) (2018). Iowa Environmental Mesonet. Available: Articles #5, Presentations 12, 16 for wake interaction, merging, meandering. Article #2 Deep array effect and atmospheric stability, Article #5 momentum transport, Presentation #17 wakes in veer and wind direction change. Article #2 on effect of surface roughness “Wind speed profiles from the IAO indicate a “surface condition bias” of 7% lower wind speed over a mixed corn/grass surface compared to a mixed corn and native vegetation surface. This translates to a 20% reduction in power for a mean 120-m wind speed of 6 m/s.”

Walton, R. A., E. S. Takle, and W. A. Gallus, Jr., 2014: Characteristics of 50 – 200 m winds and temperatures derived from an Iowa tall tower network. J. Appl. Meteor. Climatol., 53, 2387-2393. doi: Wake impact from veer and wind direction change

Appendix - Saved Comments

Andy Clifton, WindForS: Pat, I like the science goals as they are laid out, but I respectfully suggest that it might help to differentiate between basic or fundamental knowledge and things that are derived from that or even more observational / system level phenomena. This would help identify a set of basic physical understanding upon which everything else builds. For example, one could argue that a lot of the wake stuff boils down to "how do I measure and simulate my boundary conditions, how does the turbine interact with that flow, and what are the processes by which those conditions are propagated through the domain". Things like meandering, merging, loads, control decisions, and many other things can then be addressed from these building blocks. This approach fits well with the V&V methodology and also helps avoid the "we got it right for the wrong reasons" trap. Pat Moriarty, NREL reply: Thanks Andy - this is a good point - perhaps we can create a hierarchy of physics once we agree the list is complete. This may also simplify the list some and guide our testing approach.

Torben Mikkelsen, DTU: Dear Patrik. We at DTU Wind Energy would like to participate one way or the other with our full scale experimental SpinnerLidar measurements, maybe also with our ground based short-range and long-range WindScanners, to measure wakes in the near and far field wake region. Hi from Torben Mikkelsen DTU Wind Energy

Mark Stoelinga, Vaisla: Would like to add a bullet on the importance of available momentum aloft (i.e., above the rotor layer) on the rate of wake recovery. If the shear above the rotor is weak (or maybe even reverses), there is no momentum to replenish the wake from above, and all wake recovery must come from sides, which is also limited in large wind farms. We need more measurements of the wind profile in the several 100 m above the rotor layer.

Andy Clifton WindForS: WindForS would be very happy to discuss contributing fixed wing UAVs, 'copter-type UAV swarms, scanning and other lidar, and SCADA-based wake detection methods. We can also offer some UAV-carried turbulence measurement methods.

Andy Clifton WindForS: Re:”How do wakes influence blade, tower, and nacelle loading and acceleration on downstream turbines?” Is this a goal for the entire AWAKEN campaign including analysis afterwards, or is the goal to deploy enough sensors on turbines such that the validation data are available? WindForS would be happy to discuss contributing ad-hoc monitoring systems that do not require power or connection to the turbine systems in any way.

Andy Clifton WindForS: Re:” How is a wind plant wake defined?” Interesting question. I guess there are (like for individual turbines) different regions, for example the near (plant) wake is that area in which individual (turbine) wakes are visible, and the far wake is the region within which the wake manifests as a general deficit. There is some work ongoing in this area by Emeis, Bange, Siederslieben, and many others. Also, it would make sense to me to have a flexible definition depending on purpose: in some situations the "edge of the wake" would be that area where there is a velocity deficit that leads to a drop in energy produced by a turbine, while in other cases it might be better related to loads. So, avoid the "one size fits all" definition.

Andy Clifton WindForS: WindForS would be keen to discuss using the WINSENT complex terrain facility in Germany as a test case for this, potentially exchanging field and wind tunnel data or model chains.

Andy Clifton WindForS Re: “Relationship between wave height and period on tower loads (and platform/anchor system loads for FOWTs), on the flow within the plant, on wake recovery…” WindForS would be happy to discuss contributing tools and experience in this area.

Andy Clifton WindForS Re: ”Should we revisit how we model (or not model...) offshore roughness?” Yes. Essential as this a boundary condition that is almost universally accepted to be pretty poor... See e.g. work that was done for FOA 414 by NREL and NCAR (Della Monache, Hacker, Clifton, et al.), and also ongoing studies around FINO and other platforms by many European groups including WindForS, DTU, ForWind, Fraunhofer, DEWI. Happy to provide more details if required.

Andy Clifton WindForS Re: “What are the consequences of using equivalent neutral winds (satellite product) instead of stability-dependent winds when estimating vertical wind profiles and extrapolating low wind speeds to nominal hub heights?” Would prefer to see this approached as an application issue; what happens when we reduce the fidelity of input data, boundary conditions, and model physics so as to make models run faster or in areas where we have less data available? Can we still make appropriate design, controls, or operational decisions?