COMPARISON OF TURBULENCE INTEGRAL LENGTH SCALE DETERMINATION METHODS

Abstract

Grids at the inlet of the wind-tunnel test section have been commonly used to create turbulent conditions for model experiments concerning civil engineering applications. The parameters of the generated turbulence depend on the geometry of the grid and the distance from the grid to the measurement position. This work presents the experimental results of the grid-generated turbulence and a comparative analysis of an applicability of the von Kármán and Fichtl–McVehil autospectra models and autocorrelation methods (zero-crossing and exponential) for calculation of integral turbulence length scales. A square mesh array of rectangular bars was tested in a closed-circuit low velocity wind tunnel. The initial free stream turbulence was Iu = 0.75%. The measurements were carried out using the X-wire CTA probe installed at distances ranging from 1.35 m to 8.4 m downstream of the tested grid at velocities from 1.5 m/s to 9 m/s. The fitting suitability of various spectral models depends on the inflow velocity and turbulence intensity. Due to a greater flexibility and quantity of fitting coefficients, the Fichtl–McVehil model fits better in the area around the spectral peak at low flow velocity up to 3.4 m/s, while von Kármán models more accurately represent the measured spectrum at higher flow velocities. The methods based on using the correlation coefficients are equally suitable for flows at all measured velocities and turbulence intensities. The exponential method yields more stable distance-toscale characteristics with smaller deviations.