University:
Boston University
Course:
ENG ME 302 | Engineering Mechanics II
Academic year:

2019
Views:

175

Pages:

13
Author:

Clint W.

Results from LBM VLES analysis and usage in low-order Circumferential mean Axial mean APPROACH Turbulence length scale Turbulence intensity Results from LBM VLES analysis and usage in low-order 0.06 Circumferential mean Axial mean Nondimensional turbulence intensity 0.055 0.05 0.045 0.04 LBM CFD4 Turbulence intensity 0.035 0.03 0.025 0.02 0.015 0.5 0.6 0.7 0.8 Nondimensional radius TAKEOFF Nondimensional longitudinal turbulence length scale 0.04 LBM CFD4 0.035 0.03 0.025 Turbulence length scale 0.02 0.015 0.01 0.5 0.6 0.7 0.8 Nondimensional radius 0.9 1 0.9 1 Determination of turbulence isotropy LBM experiment 2 2 u s /u u u s/u u u s /u t u s/u t 1.5 1.5 u u /u t ratio ratio u u /u t 1 1 0.5 0.5 0 0 6 8 Percent span 10 5 6 7 8 radial loc (ft) 9 10 11 Stator data analysis Original goal was to use LBM VLES data to check stator response Logue/Atassi had developed 3D cascade response and shown response based on mode Difficulty is that surface pressure is due to both acoustic and hydrodynamic disturbances Method to separate cannot be used on a point-by-point basis Brian Tester is helping with some analysis and we can do more than is possible with HW but… Lengthscale from half depth wake width Approach case 0.6 0.5 Sation 2 Stationary Station 2 Wake width * 0.25 Longitudinal lengthscale (in) Longitudinal lengthscale (in) 0.6 0.4 0.3 0.2 0.1 0 20 40 60 Percent span 80 0.5 HW Stationary LBM Stationary LBM Wake width * 0.25 0.4 0.3 0.2 0.1 0 20 40 60 Percent span 80 Factor is not known across rotors Using same factor determined by experiment does not give same agreement in LBM In all cases, determine “half depth wake width” is murky and not easily automated 1 Wake evolution and length scale using separation distance Originally we planned to consider both of these things using the LBM VLES data Proven difficult We believe because of required averaging, saving, then extracting the data again For instance: If one just considers different “linecuts” extracted from the volume file 0.035 New method 450 0.03 400 0.025 (ft) 350 300 0.02 250 0.015 200 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.01 5 6 7 8 r (in) 9 10 As you follow a wake or place probes in a streamwise direction or even axial direction, the method doesn’t work well. Demonstrated this previously and don’t have work around 11 ACAT1 analysis Sketch of AneCom Aerotest (ACAT) Universal Fan Facility For Acoustics (UFFA) Rig Figure from Ricardo Blázquez(2019) ACAT1 analysis Inputs taken from Guérin (2020) ACAT1 analysis Guérin (2020) ACAT1 analysis : our results : compared to Hanson & Posson ACAT1 analysis : our results : issue of singular frequencies Summary LBM VLES seems to give no advantage for finding the input data for low-order Advantage is in more direct calculation Length scale still not nailed down Validating these calculations is tough Low-order method still gives reasonable trends ML showing promise for getting input quickly, much to be done still Acoustic kz/k range ‘k’ dashed vertical lines at Lorentz = ±1 Non-acoustic kz/k indicated ‘m’ & ‘r’ - convection at 1/Mz and 1/(0.67*Mz)