Figure 1 - uploaded by Christopher J. Bahr
Content may be subject to copyright.
Source publication
Microphone array processing algorithms often assume straight-line source-to-observer wave propagation. However, when the microphone array is placed outside an open-jet test section, the presence of the shear layer refracts the acoustic waves and causes the wave propagation times to vary from a free-space model. With a known source location in space...
Contexts in source publication
Context 1
... tunnel testing is critical for the investigation of aerodynamic and aeroacoustic characteristics of aircraft components. An open-jet wind tunnel in an anechoic chamber, such as the University of Florida Aeroacoustic Flow Facility (UFAFF) depicted in Figure 1, is often used to reduce ambient noise and minimize internal reflections over a range of frequencies. This is because the open-jet test section is free of the rigid boundaries found in the typical closed-wall arrangement. ...
Context 2
... center of the wire noise peak is approximately 37 kHz, corresponding to a Strouhal number of 0.20 based on wire diameter. Figure 10 shows the ensemble-averaged acoustic pulse spectrum compared to ensemble-averaged background noise. ...
Context 3
... the data acquired with a quiescent test section are checked for quality. An example microphone time series of a laser pulse with no flow in an empty test section is shown in Figure 11. The peak fluctuation of 50 Pa is within the linear range of the B&K microphone at ~ 128 dB (ref 20 µPa). ...
Context 4
... prior experience with the electrets, minimal harmonic distortion issues are expected at such a peak SPL. The corresponding beam map for this laser pulse, generated using the conventional frequency domain beamformer, is shown at 5 kHz in Figure 12. The source location, determined using the methods previously stated, is marked with a white 'x' within a black circle and taken as the center of the peak region in the beam map for subsequent use in source location estimation. ...
Context 5
... this case, an array data set is available for the NACA 63-215 mod-B model using a small aperture array positioned 1 m from the airfoil trailing edge [12]. As shown in Figure 13, the trailing edge noise source is shifted upstream onto the trailing edge through application of Amiet's shear layer correction described earlier. Specifically, the corrected time delays are applied to the data, but shear layer amplitude corrections are not applied or considered. ...
Context 6
... established that the shear layer correction technique works properly for an aeroacoustic source in a qualitative sense, the laser pulse data are now evaluated in a slightly more quantitative manner. In Figure 14 beam maps of the ensemble-averaged data are shown for outer array measurements at several frequencies for the case of an empty test section with no sidewalls at a flow Mach number of 0.17. Actual source locations are again shown as an overlay of a white 'x' with a black circle. ...
Context 7
... both 5.008 kHz and 10 kHz the shear layer correction appears to properly shift the predicted source location to its true location. Figure 15 shows beam maps of the ensemble- averaged data for inner array measurements at the same conditions as the previous figure for outer array measurements, although different frequencies appropriate for the inner array are considered. For these data, all of the evaluated beam maps qualitatively identify the correct source location when the standard shear layer correction method is applied. ...
Context 8
... cross-spectral phase angle is then determined by unwrapping the phase angle using this reference line. As shown in Figure 16, this reference line is necessary for phase unwrapping with these signals due to phase oscillations. A Bisquare robust line-fitting method is then used in MATLAB to determine the slope of the resultant regression line and its 95% confidence interval. ...
Context 9
... uncertainties vary from case to case, and are computed separately for each run. Figure 16. Phase angle unwrapping for outer array microphone in UFAFF empty test section at Mach 0.17, (a) with standard phase unwrapping method and (b) with the cross- correlation delay as an estimate. ...
Context 10
... normalized change in ray length is given in Equation (22), while the normalized change in propagation time is given in Equation (23). The case shown in Figure 17 is plotted again in Figure 18 with the normalized ray lengths and propagation times. With these plot parameters, interpretation of results proves easier. ...
Context 11
... normalized change in ray length is given in Equation (22), while the normalized change in propagation time is given in Equation (23). The case shown in Figure 17 is plotted again in Figure 18 with the normalized ray lengths and propagation times. With these plot parameters, interpretation of results proves easier. ...
Context 12
... present results in a more compact manner, the normalized change in ray length is plotted versus the normalized change in propagation time in subsequent figures. The first data set evaluated is the inner array data for the empty test section configuration with sidewalls, shown in Figure 19(a) through (f). The shapes of the theoretical and experimental data curves match in all the subplots, and while nominal data show offset between experimental and theoretical results, this offset is accounted for by uncertainty in all but the highest flow Mach number. ...
Context 13
... such, the airfoil wake data is evaluated. Figure 21 shows the inner array measurements for the airfoil-based trailing-edge noise experiments. For the most part, the data at higher Mach numbers follow the trend, although the Mach 0.17 data show more deviation from theory than in the empty test section case. ...
Citations
... The entries of ( ) are defined as ( ) = 0 exp , where = − 2 is the Euclidean distance separating a source-receiver pair and 0 is a reference distance. The free-space propagation time = / is used in this work with as the speed of sound in air, but could be modified to account for flow effects as in [14]. ...
... The only portion of DAMAS that depends on the number of receivers is the computation of , which has time complexity O 2 2 when computed via Eq. (14). A simultaneous form of this equation is developed to reduce computation time. ...
... The grid resolution was 0.1 in each direction. Amiet's shear layer correction method [37] was applied following the procedure described by Bahr et al. [38]. TE and LE noise were determined with the source power integration (SPI) technique [39]. ...
... P r e p r i n t n o t p e e r r e v i e w e d More details of the algorithm used are available in ref. [43]. Amiet's shear layer correction method [47] was applied following the procedure described by Bahr et al. [48]. The radiated LE noise was isolated using the source power integration (SPI) technique [49]. ...
... When the shear layer is thin and approximately planar the flow can indeed be modelled as a vortex sheet, e.g., in the correction methodology of Amiet [29,30]. The results of Amiet were confirmed by experimental work performed by Plumbee [31], Ahuja [32], and Bahr [33]. In addition to experimental work in a 0.76 m × 1.07 m wind tunnel Plumbee also performed a numerical study to assert that the influence of the shear layer thickness on the refraction angle was within 1 • , and thus neglectable in the academically scaled wind tunnel. ...
... The solution to Amiet's theory as presented by Bahr [33] is found by solving two equations simultaneously: ...
Localization of aeroacoustic sound sources in open jet wind tunnel experiments requires an accurate prediction of the acoustic propagation time. Most conventional predictions use either a ray-tracer, coupled with a modelled continuous velocity field, or use ray diffraction and a discretization of the velocity field by means of vortex sheets. In this work a novel method is proposed in which the continuous velocity field is discretized into blocks of constant velocity separated by velocity discontinuities, thus removing the requirement for the velocity to be parallel to the surface that separates the blocks. The acoustic ray is solved by minimization of the acoustic propagation time. The computational effort is low compared to ray-tracing methods while maintaining an improvement in accuracy compared to methodologies using vortex sheets. A specific continuous velocity field is derived that models a self-similar shear layer expanding asymmetrically from a rectangular nozzle. Subsequently, this velocity field is discretized to compute the acoustic rays. Experimental results with a loudspeaker source placed in the open jet of a large industrial wind tunnel showed a decrease in source localization uncertainty compared to techniques based on vortex sheets. This is attributed to the inclusion of the shear layer slanting.
... Shear-layer refraction occurs as well as coherence loss between microphone pairs. Fortunately, the shear layer refraction [30,31,32] and coherence loss [30,33,34,35] mechanisms have nowadays been widely studied and correction methodologies have been proposed and validated. Aerodynamic corrections in the OTS are typically more severe compared to the CTS, especially 75 when the lift generated by a wind tunnel model is relatively large compared to the size of the wind tunnel test section [36]. ...
... Flow-induced transmission loss was applied using the model in [37]. Moreover, a refraction correction was applied following the procedure in [32]. ...
... De-reverberation techniques were not applied. The refraction correction from Bahr et al. [32] was also applied here. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 ...
This paper assesses the comparability of aeroacoustic slat noise measurement in an open-jet, a hard-wall, and a hybrid test section configuration. The 30P30N common research model was tested in the Aeroacoustic Wind Tunnel of the University of Twente at a chord-based Reynolds number up to 1⋅106 and free-stream Mach number 0.15. Pressure distribution measurements were used to determine the aerodynamic similar condition of the flow in the slat cove. These conditions are subsequently validated by time-averaged 2D PIV measurements. The experiments show that general aspects of the flow in the slat cove are similar when the Cp distribution around the leading-edge slat element in each test section configuration is comparable. Overall, the shear layer path and reattachment point are in close agreement, showing that the aeroacoustic noise production mechanisms are comparable. Microphone phased array measurements were subsequently conducted to compare the far-field noise characteristics originating from the slat in each test section configuration. A framework is proposed to correct the acoustic measurements of each test section configuration to a standard free-field condition. Transmission loss corrections are required for the microphone measurements in the hard-wall and hybrid test section configurations, whereas a coherence loss model was validated and applied to correct the microphone measurements in the open-jet. The measured noise spectra in each test section configuration show differences up to approximately 5 dB. In general the differences were found to be smaller. The largest differences are seen for small angles of attack with noise levels in the open-jet test section showing considerably higher values than the hard-wall test section and hybrid test section results. In addition, the behavior of the vortex shedding hump in the high frequency range is different in the open-jet measurements, which could be related to the larger noise level differences at small angles of attack.
... The time-variant delay, i.e., the deviation from the average, is due to the turbulent shear layer. When the shear layer is thin compared to the wavelength the time-invariant delay time is predicted by the method of Amiet [9,10]. In this case the influence of the turbulence can be neglected. ...
Acoustic imaging methods are used to detect and quantify aerodynamic noise sources in aeroacoustic experiments. In wind tunnels with an open jet test section the sound wave propagating from an aeroacoustic source to a microphone array is distorted by the shear layer in the propagation path. The velocity fluctuations caused by the turbulence in the shear layer continuously alter the propagation time between source and observer causing reduction of the coherence between two microphone signals. The reduction in coherence is stochastic and dependents on the relative position between source and microphone. This leads to blurred acoustic source images with reduced peak Sound Pressure Level. This paper presents the Acoustic Lucky Imaging methodology, based on a methodology in astronomy to correct for image distortion due to turbulence present in Earth’s atmosphere. Furthermore, a model for coherence loss is derived to explain the loss of acoustic image resolution and allow for a-priori estimates on acoustic image blurring. The methodology is applied to an acoustic data set obtained in the open jet of a large industrial wind tunnel. It is shown that the presented methodology can restore a loss of 6 dB peak Sound Pressure Level by 3 dB and increase the resolution of the acoustic image at 8 kHz, a wind tunnel speed of 34 m⋅s⁻¹, and a shear layer thickness of ∼1 m.
... Ce retard n'est plus simplement considéré comme proportionnelà la distance entre les deux points, et dépend désormais aussi du nombre de Mach de l'écoulement, et de la position de la couche de cisaillement. La méthode développée par Amiet est décrite (et validée expérimentalement) par Bahr et al. [5]. ...
Cette étude expérimentale s'intéresse au bruit tonal de bord de fuite sur un profil NACA 0012, et en particulier aux effets de l'inhomogénéité en envergure de l'écoulement sur les sources de bruit. L'originalité de l'approche consiste dans la mesure simultanée du bruit rayonné par antennerie acoustique et des champs de vitesse dans le sillage par Vélocimétrie par Images de Particules (PIV) résolue en temps (fréquence de 20 kHz afin de résoudre les échelles de temps acoustiques).Une expérience préliminaire retrouve, comme dans la bibliographie, des résultats de PIV altérés pour des fréquences inférieures à la limite de Shannon, et permet d'identifier l'allure des distorsions liées à cette erreur. La précision de la synchronisation entre le signal focalisé par formation de voies temporelle et les champs de vitesse est ensuite évaluée, et est de l'ordre du centième de milliseconde.La mesure acoustique révèle un bruit tonal (constitué d'un ou plusieurs pics fréquentiels) pour des conditions d'écoulement en accord avec la littérature. La formation de voies permet alors d'identifier des zones sources le long de l'envergure, distinctes pour chaque fréquence. Cela constitue un apport intéressant car les modèles théoriques utilisés pour décrire la formation de bruit tonal au bord de fuite sont bidimensionnels.L'analyse fréquentielle des champs de vitesse dans un plan parallèle au profil d'aile indique un détachement tourbillonnaire variable en fonction de la position en envergure, avec les mêmes fréquences dans les mêmes zones que les sources de bruit identifiées. Il est aussi montré que les zones sources de bruit sont celles où la cohérence transverse du détachement tourbillonnaire présente un maximum. Une mesure dans le plan perpendiculaire à l'aile permet le suivi temporel des tourbillons détachés au bord de fuite, et une comparaison entre le signal de vorticité et le signal acoustique focalisé sur la zone de mesure aboutit à un taux de corrélation satisfaisant, de l'ordre de 60%.
... In the open jet, sound waves are convected with the flow and propagate through the shear layer of the jet where they are refracted. Amiet's shear layer correction method [44] is applied to account for this effect following the procedure described by Bahr et al. [45]. We consider that the coherence loss induced by the shear layer is negligible for the configuration used in this study. ...
Open-jet, hard-wall, and hybrid test section configurations are typically used in wind tunnel tests, aiming to represent the ideal free-flight condition. Each test section type requires special adaptations of the experimental hardware and specific correction methodologies that account for systematic measurement errors. Differently from previous research, this paper investigates the comparability of measurements performed in three test section configurations in the same wind tunnel. With this approach, systematic errors related to the facility or boundary-layer tripping methodology are minimized. Basic 2D aerodynamic boundary corrections are evaluated using a DU97W300 airfoil. The corrected lift curve collapsed well while differences in stall behavior and Cp distribution at larger angles of attack are nonnegligible. The trailing-edge noise produced by a NACA-0012, NACA-0018, and NACA-63018 served as reference aeroacoustic sources in this comparability analysis. Moreover, the noise reduction from trailing-edge serrations served as an additional reference for the comparability of relative noise levels. The chord-based Reynolds number of the experiments ranged from 260,000 to 660,000. Unsteady wall pressure measurements performed near the trailing-edge provided a reference for the aeroacoustic noise source terms, demonstrating a negligible change between the different test section types. The applied experimental hardware and acoustic corrections yielded aeroacoustic sources of comparable absolute far-field noise level in the three test section types within ±1−3 dB. The relative noise levels are comparable within ±1−3 dB. These results show that aeroacoustic measurements of trailing-edge noise performed in different test section configurations are equivalent, provided that adequate experimental and post-processing methodologies are systematically implemented.
... In the open-jet test section we correct for the convection of sound waves within the jet and refraction of sound waves by the shear layer is considered using Amiet's shear layer correction [9]. This method also considers the difference in the speed of sound inside and outside of the jet using the procedure described in [26]. The shear-layer coherence loss effect on the microphone phased array is considered but we assume that coherence loss is negligible since we measure at low frequencies (<5 kHz) and a low free-stream Mach number. ...
View Video Presentation: https://doi.org/10.2514/6.2021-2253.vid We experimentally investigate the comparability of airfoil trailing-edge noise measurements in an open-jet, hard-wall and hybrid test section wind tunnel in a set of benchmark exercises. We used three different airfoils, namely a NACA-0012, NACA-0018 and NACA-63018 with 0.2 m chord and 0.7 m span. We focus on the results for zero angle of attack and a free-stream velocity of 25 and 30 m/s giving a chord-based Reynolds number of 330,000 and 400,000, respectively. All measurements were conducted in the aeroacoustic wind tunnel facility of the University of Twente. This allows us to vary the test section type while keeping other parameters constant. We measured the wall pressure fluctuations near the trailing-edge on all airfoils which shows identical source terms of the trailing-edge noise mechanism when performing measurements using the same conditions in all test section types. Both the wall pressure spectrum and spanwise coherence remain constant regardless of test section type. Far-field noise is measured by a microphone phased array, and spectra are determined by using the source power integration technique of conventional beamforming maps. In the closed test section, wall-mounted microphones are recessed behind a stretched Kevlar cloth to reduce boundary layer self-noise and to improve the signal-to-noise ratio. A far-field microphone phased array is used in the open-jet and hybrid test section. Absolute noise levels of the baseline airfoils are compared and show good overall agreement of the noise levels and spectral shape within approximately 2 dB. We found an exception at higher frequencies in the closed test section where the noise level was considerably lower than the noise measured in the open-jet and hybrid test sections. We also compared the relative noise levels by evaluating the noise reduction caused by add-on trailing-edge serrations on each airfoil. The noise reduction measured in each test section was identical, within approximately 2 dB in the entire frequency range. The measured differences are typical for the amount of uncertainty that can be expected in microphone phased array measurements. The results presented provide a valuable methodology and dataset for benchmark exercises of hybrid wind tunnels and airfoil trailing-edge noise measurement and prediction.
... for . The solution due to Amiet's theory (as presented in Bahr [3]) is found by solving two equations simultaneously. Both are repeated here as ...
... The effective speed of sound 0 equals Eq. 14 presented by Bahr [3]. ...
View Video Presentation: https://doi.org/10.2514/6.2021-2130.vid Acoustic imaging methods like beamforming rely on the propagation delay between source and observer. For open jet measurements the propagation delay is computed by a method proposed by Amiet, which has been used successfully for acoustic wind tunnels since its first introduction. The sound waves are assumed to refracted on a planar velocity discontinuity - modeling the shear layer - which combined with ray theory yields the propagation path. Amiet’s method is valid when the shear layer thickness may be neglected and the shear layer is approximately planar. In this paper the Amiet correction methodology is extended to alleviate these requirements, based on a discretization of the flow in Mach iso-contours. The rays are defined as the paths that minimize the travel time between source and observer by use of Fermat’s principle. This establishes an easy mathematical optimization process. Two specific cases considered are common to all wind tunnels: (i) an expanding - slightly slanted - shear layer interface and (ii) a rounded jet core cross sectional shape. A self-similar shear layer flow model by Görtler, and an adaptation of this model that that accounts for a rounded jet core shape are used to describe the jet flow. The results of the extended model are compared to a ray-tracing solution, which is used to evaluated accuracy and the gain in computational effort. This comparison showed good agreement, while having a computational effort of two orders lower than the raytracing code. The extended model is particularly useful in large industrial wind tunnels which have thick shear layers (up to 1 m) and microphone arrays placed with an offset from the wind tunnel axis. An experimental evaluation of the method, applied to a known speaker source in a large industrial wind tunnel, shows that source location accuracy could be improved up to 6 cm.
