SWiFT Site Atmospheric Characterization

/, Energy, Renewable Energy, SWIFT, Wind Energy, Wind News/SWiFT Site Atmospheric Characterization

SWiFT Site Atmospheric Characterization

By | 2016-12-02T18:47:48+00:00 March 24th, 2016|Analysis, Energy, Renewable Energy, SWIFT, Wind Energy, Wind News|Comments Off on SWiFT Site Atmospheric Characterization

Site atmospheric conditions have been analyzed by members of Sandia’s Wind Energy Technologies Department to characterize the magnitude and frequency of atmospheric conditions at SWiFT, the experimental wind farm in Lubbock, TX. The analysis will inform experimental campaign planning at SWiFT, aid in determining common conditions for model simulation inputs. The site conditions at the lower height of the SWiFT scale turbines and provide a benchmark for comparison to site conditions at higher height utility wind farms. Two years of data from Texas Tech University’s 200m meteorological tower have been analyzed for bulk atmospheric characteristics and distributions, temporal dependency within the day and on the year, and correlations of atmospheric conditions.

Figure 1. Atmospheric variables and their probability distributions

At the 32 m hub height SWiFT scale, the site has an average wind speed of 6.8 m/s and average turbulence intensity of 11.9% over the SWiFT turbine operational wind speed range. Knowing this wind speed average is valuable for testing at SWiFT and means the National Rotor Testbed (NRT) rotor will operate at its Region-II design point for about 50% of its operation. The NRT rotor is a functionally scaled rotor to reproduce a specific utility wind turbine wake generation. Low turbulence intensity levels between 3% and 10% at the site are common which means test campaigns approximating offshore wind turbine wakes can be replicated at SWiFT with some regularity. The distribution of the atmospheric quantities is equally important to the averages. Example probability distributions for the velocity profile shear exponent and the air density at the site are shown in Figure 1. For estimating velocity profiles, a shear exponent of 0.2 is recommended by standards as an average representative value for any site (and is true at the SWiFT site). The average does not represent the velocity profile time history well, however, as shown in Figure 1. The distribution of the shear exponent has implications for wind turbine designers as the high shear cases can produce a significant load imbalance. The distribution of the air density at the SWiFT site reveals a significant +/- 10% variance from the mean, which directly scales the wind turbine power production +/- 10%.

Figure 2. Average day wind speed trends

The temporal trends of the atmospheric conditions will determine when specific experimental campaigns need to be performed and for how long. Figure 2 shows the average day wind speed history for the 10 meteorological tower stations. As noted by the spread of velocity between the stations, the overnight period has the strongest velocity profile which transitions to a turbulent profile as the sun heats the surface and the atmospheric boundary layer becomes unstable. Interestingly, the SWiFT height, 32.5 m, has a mostly flat trend of wind speed with time of day due to where it sits within the boundary layer. This is significant as it means that varying levels of turbulence and shear can be achieved at the same wind speed within Region-II operation, on the average.

Figure 3. Average year atmospheric stability probability

The SWiFT site has a stable atmospheric boundary layer over 50% of the time, unstable below 40%, and has a near-neutral stability of between 5-10%.  These results stress the importance of capturing stability classes in wind farm models and not simply assuming near-neutral stability, which is the simplest to model.  The trend of atmospheric stability for the average year is shown in Figure 3.  As would be expected, the winter months contain the highest frequency of stable cases, and summer the highest frequency of unstable cases where days are longer and there is less cloud cover.

Figure 4. Probability distribution of the shear exponent

Many atmospheric quantities correlate with each other, and this was analyzed for some of the more significant quantities of interest to wind energy.  For example, the probability distribution of the shear exponent is shown again in Figure 4, but now with stacked bar graphs representing the atmospheric stability.  This plot reveals that the low shear occurrences are mostly comprised of unstable atmospheric conditions, and that high shear is nearly entirely comprised of stable boundary layer conditions.  Neutral stability cases, where there is no vertical heat flux, occur at an average velocity profile shear exponent of 0.1 and should not be expected to occur beyond a shear exponent of 0.2.


The full report for this analysis can be downloaded from SWiFT Site Atmospheric Characterization.