Academia.eduAcademia.edu

Figure 9 – uploaded by Thomas Carolus

See full PDFdownloadDownload figure
Figure 9: Trailing edge noise extraction technique for the untripped (left) and tripped (right) airfoil; top: Cross-spectra from M 1 and M2, bottom: Extracted trailing edge noise in SPL (Af = 3.125 Hz) emission, accompanied by a strong decrease in the signal-to-noise ratio. This leads to an almost unidentifiable trailing edge noise contribution, which has also been found by other investigators, cf. Yu and JosHI [25]. Accordingly, the signals from microphones M1 and M2 have been used for a correlation analysis to extract the noise that is radiated from the trailing edge of the tripped airfoil. The correlation and filtering technique of BLAKE and LYNCH [26] has been applied. It is assumed tha trailing edge noise is of dipole character, with a phase of ©j2 = 7 between the signals reaching microphone M1 and M2 ([21], p. 490). Shear layer refraction effects [27] have been neglected in the analysis. In a first step, the cross-spectrum G)2 between M1 and M2 is determined, with and withou the airfoil installed in the mock-up. The spectra are compared, frequency by frequency. If the correlated airfoil noise for one frequency bin is less that 3 dB higher than the correlated background noise, this frequency bin is no longer considered to be assignable to trailing edge noise. For differences above 3 dB, the magnitude of the cross-spectrum background noise is subtracted from the magnitude of the cross-spectrum airfoil noise (which includes contributions from the airfoil and the facility background noise) for the frequency bin in consideration. With this, correlated background noise is eliminated and only correlated airfoil noise remains. For the frequency bins not rejected, the phase angle has been computed and if this value is in the range of 3/4 z< |O| < 5/4 z, it is assumed hat the trailing edge noise mechanism with the known phase of z is the responsible source for this. Frequency bins outside this range are rejected. The result of this analysis in terms of the SPL is presented in Figure 9, for the tripped airfoil, and for comparison purposes, also for the untripped 9girfoil.

Figure 9 Trailing edge noise extraction technique for the untripped (left) and tripped (right) airfoil; top: Cross-spectra from M 1 and M2, bottom: Extracted trailing edge noise in SPL (Af = 3.125 Hz) emission, accompanied by a strong decrease in the signal-to-noise ratio. This leads to an almost unidentifiable trailing edge noise contribution, which has also been found by other investigators, cf. Yu and JosHI [25]. Accordingly, the signals from microphones M1 and M2 have been used for a correlation analysis to extract the noise that is radiated from the trailing edge of the tripped airfoil. The correlation and filtering technique of BLAKE and LYNCH [26] has been applied. It is assumed tha trailing edge noise is of dipole character, with a phase of ©j2 = 7 between the signals reaching microphone M1 and M2 ([21], p. 490). Shear layer refraction effects [27] have been neglected in the analysis. In a first step, the cross-spectrum G)2 between M1 and M2 is determined, with and withou the airfoil installed in the mock-up. The spectra are compared, frequency by frequency. If the correlated airfoil noise for one frequency bin is less that 3 dB higher than the correlated background noise, this frequency bin is no longer considered to be assignable to trailing edge noise. For differences above 3 dB, the magnitude of the cross-spectrum background noise is subtracted from the magnitude of the cross-spectrum airfoil noise (which includes contributions from the airfoil and the facility background noise) for the frequency bin in consideration. With this, correlated background noise is eliminated and only correlated airfoil noise remains. For the frequency bins not rejected, the phase angle has been computed and if this value is in the range of 3/4 z< |O| < 5/4 z, it is assumed hat the trailing edge noise mechanism with the known phase of z is the responsible source for this. Frequency bins outside this range are rejected. The result of this analysis in terms of the SPL is presented in Figure 9, for the tripped airfoil, and for comparison purposes, also for the untripped 9girfoil.

Related Figures (9)

adverse pressure gradient occurring at high curvature. Two interchangeable square nozzles have been manufactured with exit-plane dimensions of 133x133 mm’ and 180x180 mm’, i.e. contraction ratios of 9 and 4.9 with relative contraction lengths of 0.75 and 0.82, respectively. The chosen nozzle contour is of minimum curvature; it is described by two intertwining radii which meet at the matching point, located in the middle between nozzle entrance and exit. Figure 3 shows a scheme of the nozzle, along with the implemented final design. The actual nozzle exit-side (at Do) is extended by Ax = 0.075D, in length to obtain smoother exit conditions for the flow. AEROACOUSTIC AND AERODYNAMIC CHARACTERIZATION
adverse pressure gradient occurring at high curvature. Two interchangeable square nozzles have been manufactured with exit-plane dimensions of 133x133 mm’ and 180x180 mm’, i.e. contraction ratios of 9 and 4.9 with relative contraction lengths of 0.75 and 0.82, respectively. The chosen nozzle contour is of minimum curvature; it is described by two intertwining radii which meet at the matching point, located in the middle between nozzle entrance and exit. Figure 3 shows a scheme of the nozzle, along with the implemented final design. The actual nozzle exit-side (at Do) is extended by Ax = 0.075D, in length to obtain smoother exit conditions for the flow. AEROACOUSTIC AND AERODYNAMIC CHARACTERIZATION
A single 2” B&K microphone (model 4190, 1 m from the middle of the tunnel exit, under 45° to the center line) was used to measure the SPL (sampling rate: 25.6 kHz, Af = 3.125 Hz) at the nozzle exit. Measurements were made with the empty muffler casing installed, the fully assembled muffler, and then the muffler with the small nozzle, each at the driving fan’s max. operating conditions. Figure 4 (left) shows the results in comparison to the background noise level.
A single 2” B&K microphone (model 4190, 1 m from the middle of the tunnel exit, under 45° to the center line) was used to measure the SPL (sampling rate: 25.6 kHz, Af = 3.125 Hz) at the nozzle exit. Measurements were made with the empty muffler casing installed, the fully assembled muffler, and then the muffler with the small nozzle, each at the driving fan’s max. operating conditions. Figure 4 (left) shows the results in comparison to the background noise level.
The flow field has been measured with a standard a 1-D hotwire probe (DANTEC 55P11, sampling rate: 20 kHz), for both nozzle sizes and for three different velocities. Some example plots of mean and turbulent quantities for the larger nozzle are presented in Figure 5. The extent of the potential core of the jet is identifiable from both graphs: the turbulent mixing region is indicated by the increased level of turbulence intensity, as well as the decrease in the mean velocity. Profile non- uniformities, mainly at x/Do = 0, may arise from manufacturing inaccuracies of the nozzle. The bias could also be due to the influence of S-shaped bend on the flow, since the small nozzle shows quite similar velocity profiles. Overall, the minimum turbulence intensity is found to be around 0.4% in the middle of nozzle exit-plane and increases to levels of 5%-15% in the outer skirt of the potential core, where intense mixing occurs. For airfoil experiments, the streamwise location x/Dpg = 0.56 is important, since it marks the distance, where the airfoil leading edge will be located. At this distance, a turbulence intensity of around 0.36%-1% is present in the core region and the mean velocity profile is rather uniform as compared to the first plane.
The flow field has been measured with a standard a 1-D hotwire probe (DANTEC 55P11, sampling rate: 20 kHz), for both nozzle sizes and for three different velocities. Some example plots of mean and turbulent quantities for the larger nozzle are presented in Figure 5. The extent of the potential core of the jet is identifiable from both graphs: the turbulent mixing region is indicated by the increased level of turbulence intensity, as well as the decrease in the mean velocity. Profile non- uniformities, mainly at x/Do = 0, may arise from manufacturing inaccuracies of the nozzle. The bias could also be due to the influence of S-shaped bend on the flow, since the small nozzle shows quite similar velocity profiles. Overall, the minimum turbulence intensity is found to be around 0.4% in the middle of nozzle exit-plane and increases to levels of 5%-15% in the outer skirt of the potential core, where intense mixing occurs. For airfoil experiments, the streamwise location x/Dpg = 0.56 is important, since it marks the distance, where the airfoil leading edge will be located. At this distance, a turbulence intensity of around 0.36%-1% is present in the core region and the mean velocity profile is rather uniform as compared to the first plane.
Figure 6: Turbulence energy spectra (left) compared to the VON KARMAN spectrum (black lines) and autocorrelation functions (right) for sample points (bn = big nozzle, sn = small nozzle)
Figure 6: Turbulence energy spectra (left) compared to the VON KARMAN spectrum (black lines) and autocorrelation functions (right) for sample points (bn = big nozzle, sn = small nozzle)
Figure 8: Left: SPL of M2 for different test arrangements with and without model airfoil and tripping (serrated tape). Right: SPL (40 cm from the trailing edge) of the airfoil with and without tripping (1.9: mm wire) and with an artificially thickened trailing edge (1.95 mm wire attached) without tripping
Figure 8: Left: SPL of M2 for different test arrangements with and without model airfoil and tripping (serrated tape). Right: SPL (40 cm from the trailing edge) of the airfoil with and without tripping (1.9: mm wire) and with an artificially thickened trailing edge (1.95 mm wire attached) without tripping
Figure 10: Mean and turbulent velocity profiles of the airfoil wake for different downstream locations from the trailing edge. Mean profile are measured from the wake center
Figure 10: Mean and turbulent velocity profiles of the airfoil wake for different downstream locations from the trailing edge. Mean profile are measured from the wake center
Related topics:
Mechanical EngineeringClassical PhysicsManufacturing EngineeringWind Tunnel
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved