Analysis of Turbulence Profiles from Three Tall Towers: Departure from Similarity Theory in Near-Neutral and Stable Conditions

Abstract

Long-term wind and turbulence profiles were analyzed for all stability conditions at three tall, multi-level towers located at the Los Alamos National Laboratory (LANL), Rocky Flats Environmental Plant (RF), and the Boulder Atmospheric Observatory (BAO). The LANL and RF sites are located in complex terrain and the BAO is located over relatively simple terrain, but within 3 to 5 km of an abrupt 20 to 30 m increase in terrain. Results indicate that normalized turbulence parameter profiles at all three sites agree well with widely used empirical relationships during unstable conditions. During near neutral conditions, σu parameter profiles are also well behaved at all three sites while σw increases with height for complex fetch (BAO downwind of bluff, LANL, and RF) while σw remains nearly constant up to 200 m AGL at BAO with simple fetch. The σw /u* values at 10-m AGL are close to one at all sites and they increase by an order of 50% in the lowest 60 to 200 m for complex fetch and remain approximately constant in the lowest 200 m with simple fetch. During very stable conditions, typical values of σu and σ v range between 0.4 to 0.6 ms-1 and increase slightly with height while median σw values nearly double from about 0.1 to 0.2 ms-1 between the 10- and 100 to 200-m levels. A comparison of predicted with measured u* values at two of the sites shows generally good agreement over 6 stability categories. It is suggested that M-O similarity theory will usually greatly underestimate vertical diffusivity and dispersion during very stable conditions, especially at larger heights, based on idealized Kz profiles calculated from measured σw values. Finally, rules of thumb are formulated to describe departure from similarity theory during near-neutral and stable conditions.

Climate Adaptation Engineering and Risk-based Design and Management of Infrastructure

“…BS EN 61400-1 ( 2005) assumes that the longitudinal turbulence standard deviation 1 is invariant with height. This is deemed a valid assumption, having been also implemented by the design standard DNV-OS-J101 (2010) and was demonstrated to have sufficient accuracy by Bowen (2008). The code also suggests that a coherence model be specified to correlate the turbulence effects across the entire rotor.…”

Fragility analysis of steel and concrete wind turbine towers

“…The fluctuating, or turbulent, wind velocity time-histories, are generated using the Bowen (2008). T he code also suggests th a t a coherence m odel be specified to correlate th e tu rbulence effects across th e entire rotor.…”

Roughness Lengths for the Savannah River Site

“…where A, B and C are constants that have been determined to be 2.5, 1.9 and 1.25 (Panofsky and Dutton, 1984;Arya, 2001;and Bowen, 2008).…”

“…However, some studies have noted that use of  u , rather than  E yields a roughness value biased by the effects of upwind terrain changes and obstacles over many kilometers (Tielman, 1992, andBowen, 2008). Conversely, roughness based on vertical turbulence intensity is more likely to reflect a more relevant local roughness for the surrounding vegetation.…”

“…In the past, the use of  u to compute roughness has been discouraged because it reflects large-scale roughness features far upstream that are slow to decay (Tielman, 1992;Bowen, 2008 Another factor that may cause smaller roughness values from the   and  u methods is that the 15-minute averaging time for these quantities may not capture the full spectrum of relevant turbulence scales, which decreases the value of roughness they predict. Even if a one-hour average of either   and  u were used, it would include changes in mean values of wind direction and speed between each 15-min period.…”