AIAA-2006-9-275
27 Pages

AIAA-2006-9-275

Course Number: IS 275, Fall 2008

College/University: Cincinnati

Word Count: 10763

Rating:

Document Preview

44th AIAA Aerospace Sciences Meeting and Exhibit 9 - 12 January 2006, Reno, Nevada AIAA 2006-9 Broadband Shock Associated Noise Suppression by Chevrons Olaf H. Rask* and Ephraim J. Gutmark University of Cincinnati, Cincinnati, Ohio, 45221-0070 Steven Martens GE Aircraft Engines, Springdale, Ohio, 45246, USA Tests were conducted to determine the effect of chevrons on the acoustic emissions from nozzles operating...

Unformatted Document Excerpt
Coursehero >> Ohio >> Cincinnati >> IS 275

Course Hero has millions of student submitted documents similar to the one
below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.

Course Hero has millions of student submitted documents similar to the one below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.

AIAA 44th Aerospace Sciences Meeting and Exhibit 9 - 12 January 2006, Reno, Nevada AIAA 2006-9 Broadband Shock Associated Noise Suppression by Chevrons Olaf H. Rask* and Ephraim J. Gutmark University of Cincinnati, Cincinnati, Ohio, 45221-0070 Steven Martens GE Aircraft Engines, Springdale, Ohio, 45246, USA Tests were conducted to determine the effect of chevrons on the acoustic emissions from nozzles operating at underexpanded conditions. Conical (as baseline) and chevron nozzles were tested on the core (primary) stream in a coaxial nozzle configuration that simulated a high bypass ratio engine exhaust. The secondary (fan) stream was varied from quiescent condition to Mach=0.85. In a static condition, the chevron nozzle was shown to result in higher shock noise levels by 0.7 dB (OASPL). Additionally, the shock noise frequency for the chevron had shifted to higher levels. For the M=0.85 condition, the chevron nozzle was shown to result in lower shock noise levels by 2.1 dB (OASPL). As with the static case, the shock noise frequency for the chevron had shifted to higher values. For all secondary fan stream velocities, it was shown that the chevron nozzle reduced the shock cell spacing, which lead to a higher frequency for shock noise. These tests can also be considered as an approximation to the effect of free stream on fan chevrons efficiency during cruise. Nomenclature ATF a D Deq f k L Md Mj NB NPR OASPL Req SPL TKE uc X 1 = = = = = = = = = = = = = = = = = = = AeroAcoustic Test Facility at the University of Cincinnati ambient speed of sound nozzle diameter equivalent nozzle diameter center/peak frequency of broadband shock associated noise ratio of specific heats shock cell spacing nozzle design Mach number fully expanded jet Mach number Narrowband Nozzle Pressure Ratio Overall Sound Pressure Level Radial displacement, nondimensionalized with Deq Sound Pressure Level Turbulent Kinetic Energy turbulence convection velocity Axial coordinate directivity angle (180 is the direction of the jet plume) first zero of Bessel function of order 0, J0( 1)=0 I. Introduction O * VER the past several years, chevrons have been studied extensively as a means to reduce turbulent mixing noise emitted from medium to high bypass turbofan engines1-6. This work has shown that primary and Graduate Student, University of Cincinnati, AIAA Student Member Professor, Ohio Eminent Scholar, Aerospace Engineering, University of Cincinnati, AIAA Associate Fellow Senior Research Engineer, GE Aircraft Engines, Springdale, Ohio, AIAA Member 1 American Institute of Aeronautics and Astronautics Copyright 2006 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved. secondary nozzle chevrons can significantly increase the exchange of momentum from the primary stream to the secondary stream, and from the secondary stream to the free stream. The rate of exchange is controlled by two principle parameters, chevron penetration and shear velocity across the chevrons. With relatively low penetration chevrons, the level of mixing between adjacent streams is increased, leading to reduced levels of jet noise, particularly at lower frequencies. With relatively high penetration chevrons, the level of mixing between adjacent streams is significantly increased. This leads to even higher levels of low frequency jet noise reduction. However, as a consequence of the more aggressive mixing, excess turbulence is created immediately downstream of the chevron nozzle. This can result in a significant high frequency noise penalty. A similar effect has been observed with shear velocity. At lower shear velocities across a chevron nozzle, mixing is increased to a small degree and jet noise is reduced. At higher levels of shear velocity, mixing is significantly increased, resulting in much more low frequency jet noise reduction. However, as with penetration, high shear velocity can create excess turbulence immediately downstream of the nozzle. This can also result in a high frequency noise penalty. Jet noise due to the high speed flow and the turbulent mixing in the plume that is emitted from a turbofan engine is of concern to the general public only when an airplane is within several thousand feet of the ground. Above that altitude, the reduction of jet noise due to the increased observer distance and the absorptive properties of the atmosphere significantly reduce the problem. For this reason, chevron studies have focused on take-off type conditions. However, as an airplane climbs to cruise altitude, discharge conditions may become under expanded. As a result of this supersonic discharge, two additional types of jet noise are created, screech (discrete tones) and broadband shock associated noise. When the nozzle pressure ratio (NPR) across a convergent nozzle is equal to approximately 1.89, the flow discharges parallel to the axis and at sonic velocity. For these conditions and downstream of the nozzle, the static pressure of the discharge has been reduced such that it is equal to the ambient pressure. If the NPR increases above 1.89, the discharge is still at M=1 and parallel to the axis. However, the nozzle is no longer able to expand the flow to ambient pressure. In order to match the two pressures across the slip line, the flow is turned away from itself through an expansion fan. A series of expansion fans and oblique shocks are formed in the jet downstream of the nozzle until the static pressure in the jet becomes equal to the ambient conditions. The resulting structure is called a quasi periodic series of shock cells. Two such cells are illustrated in figure 17. Research concerning the noise created by imperfectly expanded jet flows surged with the expected community noise issues arising from takeoff and landing cycles of the High Speed Civil Transport8-15. Specifically, broadband shock associated noise is created as turbulence structures pass through the quasi periodic series of shock cells of an imperfectly expanded jet. This interaction with successive shock cells results in constructive/destructive interference that creates preferred frequencies depending upon the observers angle. Experimental investigations have shown that shock noise is most prominent at forward angles. The center/peak frequency increases at more aft angles. In addition, the dominance of shock noise with respect to turbulent mixing noise is reduced. Equation 1 gives the relationship between the peak frequency and the direction of radiation of broadband shock associated noise8,16. This equation is valid for a single jet exhausting into still air. f= uc u L(1 + c cos( )) a (1) The value of uc has been shown to be approximately 70% of the fully expanded jet velocity16. The typical behavior of shock noise creation can be seen from equation 1. At 90, the center frequency of shock noise is the frequency at which turbulent structures pass through the shock cells. Additionally, it is generally accepted that shock cells are not stationary17. Their interaction with turbulence causes them to vibrate back and forth slightly in the streamwise direction. This jitter increases with downstream direction and acts to increase the spectral width of shock noise. In addition, the convective velocity of turbulence structures moving through the shock cells is not uniform. Any variation in convective velocity will also increase the spectral width of shock noise18. It general, it can be seen from equation (1) that any increase in the variation of shock cell spacing or turbulence convection velocity will result in a broader spectral half width of shock associated noise. Based in part on the above theory, Tam19-20 2 American Institute of Aeronautics and Astronautics developed equations to calculate the shock associated noise of imperfectly expanded jets exhausting into still air. Comparisons with the experimental measurements by Norum et. al.11 were good, suggesting that the theories were adequate. The normalized shock cell spacing can also be estimated by using equations from Tam15. Equation (2) is valid for a single jet exhausting into still air. k +1 4 ( k 1) L = D ( M 2 1) 1 + j 1 (k 1) M 2 j 2 2 (k 1) M d 1+ 2 Md Mj (2) Some of the cited research12-14 also investigated the influence of forward flight on broadband shock associated noise. Typically, this is done by placing the supersonic discharge in a wind tunnel or a free jet. Experimental results have shown that the addition of a second stream reduces the center frequency and the spectral half width of shock noise. From a flow field point of view, the second stream reduces the shear velocity across the two streams. This reduction decreases the mixing which increases the length of the potential core and also the shock cell spacing. The increased shock cell spacing leads directly to reduced peak frequency for the shock associated noise as can be seen from equation (1). As a consequence of the reduced shear, the TKE in the shear layer is also reduced. The change in TKE may reduce the spatial jitter of the shock cells, which would lead to reduced spectral half width of shock noise. In addition, the shock noise must now pass through a moving medium (mean flow convection effect) which also tends to reduce the spectral half width21, 22. Finally, the convective velocity of turbulence in the shear layer is now approximately 70% of the fully expanded jet velocity plus 30% of the flight velocity21. With these observations, Tam22 further modified his equations to accommodate a flight velocity and obtained good comparison with experimental results. Another important shock associated narrow band noise component is the screech that can occur whenever a flow is imperfectly expanded. Noise created by turbulence/shock cell interaction propagates upstream and just outside of the jet. This noise excites the shear layer in the vicinity of the nozzle lip, creating additional turbulence, completing a feedback loop. The frequency of screech can also be estimated by equation (1) with the substitution, =0. Because screech tones are intimately associated with the nozzle lip, they are exacerbated if the nozzle has a thick trailing edge. As was discussed, a flight velocity increases the shock cell spacing. Therefore, a flight velocity would also reduce the frequency of screech. Previous experiments13 using a convergent nozzle with a fully expanded Mach number = 1.17 created a screech frequency was 9000 Hz for M=0 which dropped to 7500 Hz for M=0.40. As the noise due to turbulent mixing is reduced with higher bypass ratio engines, broadband shock at cruise conditions is emerging as the dominant noise component. Shock noise radiates predominately at upstream angles relative to the direction of flow. With respect to an airplane, this is toward the cabin upstream of the engine exhaust. Because the conditions required to create shock associated noise are present for most of a typical flight, it could have a direct and detrimental effect on passengers and crew. It would be ideal if chevrons could be used to reduce both the jet noise associated with turbulent mixing and also broadband shock associated with the under expanded condition. It is reasonable to presume that the shock cell pattern behind a chevron nozzle will dissipate faster than the shock cell pattern behind a baseline nozzle. It is also reasonable to presume that the turbulence levels immediately behind the chevron nozzle will increase compared to the levels associated with a baseline nozzle. What cannot be predicted with certainty is the effect on the broad band shock associated noise when the increased turbulence levels interact with the reduced number of shock cells. Recently, tests were conducted in a wind tunnel to assess the effect of both core and fan chevrons on shock cell noise23, 24. A 5dB variation in shock cell noise was demonstrated between different configurations of baseline and chevron nozzles. However, at certain conditions, core chevrons reduced shock noise while fan chevrons increased shock noise. At other conditions, fan chevrons reduced shock noise while core chevron increased shock noise. The primary focus of the current work is to investigate the 3 American Institute of Aeronautics and Astronautics effect of chevron nozzles on shock associated noise and to demonstrate a relationship with the flow field. In addition, discrete tones and turbulent mixing noise will be presented. II. Experimental Set-up, Procedure, and Sources of Error The Aeroacoustic Test Facility is housed in a 24 x 25 x 11 anechoic chamber with a lower cutoff frequency of ~500 Hz. The dual stream jet is a reduced scale model of the exhaust section of a modern jet engine. The pressure supply used to operate the ATF allows for an under expanded core flow and a sonic fan flow. The core flow can be heated to a maximum 250F while the fan flow is maintained at 80F. An external centerbody is located in the core nozzle. Acoustic measurements were performed using eight B&K condenser microphones, which can measure frequencies up to 100 kHz. For far-field measurements, the microphones were placed in a radial arc at 70 equivalent nozzle diameters covering angles from 70 to 150. For reference, 180 is in the direction of the jet plume. For nearfield measurements, the microphones were set in a linear array that was mounted on a computer controlled traverse. During the test, the computer alternately acquired acoustic data and moved the array until the acquisition grid was fully populated. The grid extended approximately 22 Deq in the axial direction and approximately 7 Deq in the radial direction. The grid was also rotated 7 to keep the microphones out of the jet plume. All acoustic measurements were stored as binary time records. During tests, data is displayed in real time for test control and is post processed to provide both NB and one-third octave band spectral data. Overall Sound Pressure Levels were also calculated. In this work, all acoustic data are As measured, meaning they have not been corrected to standard temperature or humidity. This work is concerned with the near field of shock associated noise. Any changes in noise due to variations in temperature and humidity are small. Measurements of static pressure measured by a Pitot probe are also presented. The probe was mounted on a computer controlled traverse and was positioned 0.1 Deq off the centerline. The off center location was used to eliminate any influence of the centerbody wake on the measurements of the shock cell separation. Ten pressure measurements were acquired for every equivalent diameter of axial distance. The number of points was considered a tradeoff between an adequate number of points to define the shock cell pattern and the air available to complete a test. A detailed discussion of the facility is available elsewhere25. The ATF primary stream was set to a NPR of 2.36, which with the convergent nozzle, yielded an under expanded jet, creating a series of quasi-periodic shock cells with a peak (fully expanded) Mach number of 1.18. The ATF secondary stream was set to NPRs of 1, 1.056, 1.187, and 1.604, which result in Mach=0, 0.28, 0.50, and 0.85. These conditions produced a range of shear velocities across the primary nozzle and could be considered as a crude simulation of static, takeoff, climb out, and cruise flight conditions relative to an under expanded fan nozzle. In order to assess the effectiveness of secondary chevrons in these conditions, the baseline primary nozzle was replaced with a low penetration chevron nozzle. This chevron nozzle was optimized for community jet noise reduction of the core stream and would likely have to be redesigned for optimal broadband shock noise reduction on the fan nozzle. When conducting tests in the ATF, significant efforts are aimed at maintaining the correct pressures that are called out on the test card. However no facility, including the ATF, has a perfect system. Far-field tests last for approximately 11 seconds. Due to the relatively short test duration, multiple data sets at the same cycle condition are typically acquired. In this way, the likelihood that at least one of them falls within the tolerances of experimental error is quite good. Shown on figure 2 are two sets of redundant data from a chevron nozzle. The secondary flow NPR is equal to 1.00 (Mach number = 0) and the intended primary NPR is 2.36. Because there is no flow in the secondary stream, any inconsistency in noise can be entirely attributed to variations in the primary flow. During post test analysis, it was discovered that the primary stream NPR varied between the two tests by 0.5%. Notice that this discrepancy is in no way reflected by the turbulent mixing noise. The two sets look almost identical. However, the difference in the discrete tone at 4000 Hz is almost 1.5 dB. The difference in shock associated noise is almost 2 dB. Shock noise (discrete and broadband) is much more sensitive to NPR than mixing noise. To emphasize this point, two sets of baseline data are shown in figure 3. For this data set, the intended secondary stream NPR was 1.604 and the intended primary stream NPR was 2.36. During post test analysis, it was discovered that the primary stream NPR varied by 0.03 %. Despite this tiny error, the shock associated noise between the two data sets differs by as much as 0.5 dB. During the same post test analysis, it was discovered that the secondary stream NPR varied by almost 2%. Despite that relatively large error, the two data sets look almost identical with respect to mixing noise. Clearly, shock associated noise is much more sensitive to NPR than turbulent mixing noise. 4 American Institute of Aeronautics and Astronautics Near-field tests take approximately 12 minutes. During near-field tests, it is possible to know the exact pressures and temperatures of the flow when microphones are recording at different locations in the mapping. Initial, nearfield test results showed that shock noise with secondary Mach = 0.28 and 0.50 was much more sensitive to variations in core NPR than shock noise with secondary Mach = 0 or 0.85. Large, discontinuous variations in SPL corresponded to small variations in core NPR. Similar variations in core NPR did not cause significant variations in the near-field mapping for secondary Mach = 0 or 0.85. Tests with less than acceptable results were redone . III. Results and Discussion A. Far-field Acoustic Shown in figure 4 are the eight measured narrow band spectra from the cycle point with Secondary Mach = 0. The individual SPL have been displaced vertically by an arbitrary amount to allow uncluttered viewing. In addition, they are labeled according to their angle of measurement where 70 is the most forward angle and 150 is the most aft angle. The characteristic hump of broadband shock associated noise can be seen at ~7500 Hz for the 70 spectra. At more aft angles, the center frequency increases and the relative dominance of shock noise with respect to turbulent mixing noise is reduced. A discrete tone at 4000 Hz for the 70 SPL marks the lower boundary of the broadband shock noise. These three observations are well known characteristics of shock noise. Equation (1) has been used to estimate the center frequency of the shock noise. Pitot probe measurements (to be discussed later) were used to experimentally determine shock cell spacing, which is required for frequency calculations. It was found that a convective velocity of 73% of fully expanded jet velocity yielded better estimates than 70%. The calculated values have been superimposed on figure 4. The estimate for 70 is approximately 1000 Hz low. However, estimates are excellent at all other angles. Figure 5 is the scale model jet noise narrowband SPL at 150 for baseline conic nozzles at the four cycle conditions. Over the majority of the spectrum, the static condition (Secondary Mach = 0) emits the highest jet noise. The inner shear layer gradient is the strongest at this condition, creating the most turbulence and noise. At this condition, there is no outer shear layer. With a secondary flow of Mach 0.28, the jet noise is reduced. This happens because the gradient across the inner shear layer is weaker, and produces less jet noise. Although there now exists an outer shear layer, it is not making sufficient noise to be heard in comparison to the inner shear layer. Accelerating the fan stream to Mach 0.50 reduces the jet noise to a minimum of the cycle points tested. This condition has further reduced the gradient of the inner shear layer and the noise it makes. The outer shear layer is stronger, but the noise created by it is still below, or at least comparable to, the noise of the inner shear layer. Further accelerating the secondary stream to Mach 0.85 increases the noise level. With this condition, the inner shear layer is further weakened and is making even less noise. However, the outer shear layer, which is stronger at this condition, now dominates. Norum and Shearin13 tested a convergent nozzle exhausting at M=1.17 (fully expanded) and simulated flight velocities from M=0 to M=0.4. The mixing noise (measured at =150 and peak frequency) dropped from approximately 99 dB at M=0 to a minimum of approximately 85 dB at M=0.35. Mixing noise increased to approximately 90dB at their testing limit of M=0.40. These results suggest that at M>0.35, the noise of the outer shear layer exceeds the noise of the inner shear layer, which is different compared to the current work. The current tests used a supersonic jet with an equivalent diameter of ~2 inches with a coaxial jet of ~4 equivalent diameter inches. The supersonic jet used by Norum and Shearin13 had a diameter of 1 inch, while the free jet had a diameter of 18 inches. With a much larger free jet to supersonic jet diameter ratio, it is expected that the outer shear layer will dominate the inner shear layer at a lower Mach number. Norum & Brown14 acquired far-field acoustic data using a 0.75 inch diameter C-D nozzle designed for perfectly expanded discharge at M=1.5. Their free jet was 12 inches in diameter. With the nozzle exhausting at design conditions, the measured level of mixing noise ( =150 and peak frequency) was 103 dB at M=0 and 101 dB at M=0.1. Mixing noise (measured at =140 and peak frequency) dropped to 97 dB for M= 0.2, 94 dB for M=0.3, 90 dB for M=0.4, and reached a minimum of 86 dB at their testing limit of M=0.5. The data from these tests would suggest that noise from the inner shear layer dominates noise from the outer shear layer at all conditions tested. This is not surprising as the supersonic jet is discharging at a much higher Mach number. It may be assumed that had the testing continued to higher coaxial velocities, a minimum in mixing noise would have been observed. Results from these two tests13,14 are consistent with the current investigation. 5 American Institute of Aeronautics and Astronautics Figure 6 shows the narrowband SPL at 90 for the baseline nozzle of all the four cycle conditions. This angle is shown because shock noise is more intense at the forward angles and it yielded good estimates for peak frequency of shock noise for M=0. The shock associated noise for all cycle points can be seen between approximately 3,000 and 10,000 Hz. Outside of that band is the turbulent mixing noise. The same trends in turbulent mixing noise that were seen at 150 are evident in this figure. The highest level of jet noise is created when the secondary stream is at M=0. Accelerating the secondary stream to Mach 0.28 reduces the jet noise. Further accelerating the secondary stream to mach 0.50 reduces jet noise to a minimum. The final secondary mach number of 0.85 increases jet noise. The shock noise level follows a similar trend with respect to secondary Mach number. Increasing the secondary flow Mach number from 0 to 0.28 only slightly reduces the shock noise. Norum & Sherin13 and Norum & Brown14 also showed very little effect of a low secondary flow on broadband shock associated noise. A further increase to Mach 0.50 results in a much lower level of shock noise. Increasing the secondary flow to Mach 0.85 significantly elevates the shock noise. Although the trends are similar, the sensitivities to secondary flow are significantly different. There is a negligible change in shock associated noise change between Mach 0 and 0.28 while the turbulent mixing noise dropped by approximately 2 dB. The changes in shock noise and turbulent mixing noise are comparable between secondary flows of Mach 0.28 and 0.50. When the secondary flow is increased to its final value, the shock associated noise increased by approximately 11 dB while turbulent mixing noise only increased by less than 4 dB. It can also be seen that the center frequency of the shock cell noise decreases with increasing secondary flow speed. This is consistent with previous experimental results12-14,17. Using equation (1), the peak frequencies have been estimated and are superimposed on the figure. In general, the agreement is good. It can be seen that the spectra become narrower with increasing secondary flow speed. This is also consistent with previous work12-14. Secondary flow reduces the shear velocity and TKE across the flow. As discussed by Tam16, any reduction in TKE reduces the spatial jitter of the shock cell pattern. This results in a narrowing of the spectral half width of the shock associated noise. Screech tone frequency can be predicted from equation 1 by using =0. Because secondary flow increases the shock cell spacing, it is expected that it will also reduce the frequency of screech. This can be seen in figure 6, where the screech frequencies for flight speeds of M=0, 0.28, and 0.50 are 4100 Hz, 3950 Hz, and 3150 Hz, respectively. Previous research13 with a similar under expanded condition generated a screech tone of approximately 9000 Hz at M=0 and a screech tone of approximately 7500 Hz at M=0.40. The SPL at 150 for the chevron nozzle at all cycle conditions are shown in figure 7. The trends in noise level with respect to secondary flow Mach number are the same that were seen with the baseline nozzle. Jet noise is reduced when the secondary stream is increased from Mach 0 to 0.28. Jet noise drops to a minimum with the secondary flow at Mach 0.50. It then increases with a secondary flow Mach number of 0.85. It can be seen that the cycle condition with no secondary flow, or the highest shear velocity across the chevron nozzle, appears to have an additional high frequency (5000 11000 Hz) component. This is not surprising in that a high frequency penalty is typically seen with increasing shear velocity. A low penetration chevron nozzle was used and is not likely to create a high frequency penalty for the other three conditions that are at lower shear velocities. The SPL at 90 for the chevron nozzle and all cycle conditions are shown in figure 8. It can be seen that the jet noise due to turbulent mixing is reduced, particularly between 1000 and 5000 Hz, when the secondary flow Mach number is increased from 0 to 0.28. Increasing the secondary flow Mach number to 0.50 reduces the mixing noise to a minimum between 1000 and 6000 Hz. Jet noise in this band increases to a maximum when the secondary flow Mach is brought up to 0.85. This trend in mixing noise with respect to secondary flow Mach number is consistent with everything already presented. The trend does not exist for high frequency (>15000 Hz) turbulent mixing noise. However, these frequencies are most likely to be affected by the high frequency penalty of chevrons. This penalty increases with shear velocity. It can be seen that the highest level is also associated with the highest shear velocity across the chevron nozzle. The levels decrease with decreasing shear velocity. The shock associated noise follows a similar trend compared to the baseline nozzle. The level with the secondary flow set to Mach 0.28 is slightly higher than the level for M=0. While this differs from the baseline nozzle, the shock noise level is very insensitive to changes in secondary flow Mach number at these low values. As with the baseline nozzle, the shock noise level of the chevron drops significantly when the secondary flow is increased to Mach 0.50. In addition, the center frequency appears to drop slightly. As with the baseline nozzle, there is a significant increase in the shock noise level with a secondary flow increase to Mach 0.85. The results presented so far show that, for the most part, general trends in shock associated noise and turbulent mixing noise with respect to secondary flow are consistent when comparing results from a baseline nozzle or a chevron nozzle. Next, the differences between baseline and chevron nozzles at specific cycle points are examined. 6 American Institute of Aeronautics and Astronautics The SPL at 150 for both baseline and chevron nozzles for all secondary flow Mach number are shown in figure 9. SPL for different secondary flow velocities have been vertically offset by an arbitrary amount to allow for easy viewing. For M=0, chevrons reduce mixing noise at the peak frequency of 1250 Hz by approximately 3.5 dB. For M=0.28, chevrons reduce jet noise at the peak frequency of 1000 Hz by approximately 3dB. For M=0.50, the reduction at peak frequency of 850 Hz is just over 2 dB. For M=0.85, chevrons reduce jet noise at peak frequency of 700 Hz by less than 1 dB. It is widely accepted that low frequency chevron noise reduction diminishes with a reduced shear velocity2-4,6. As coaxial velocity increases, the shear velocity across the jet plume drops which reduces chevron effectiveness. This is shown in the current data and is completely consistent with previous observations of chevron mechanisms. Alternatively, at M=0 (the highest shear velocity) a high frequency penalty due to the chevrons can be seen between 10,000 and 30,000 Hz. At M=0.28, excess high frequency noise due to the chevron slightly exceeds noise of the baseline nozzle at 25000 Hz. At M=0.50, excess noise created by the chevron nozzle at 25000 Hz is slightly below noise from the baseline nozzle. At M=0.85 (the lowest shear velocity), there is no excess high frequency jet noise. Previous research2-4,6 has also demonstrated that a high shear velocity across a chevron nozzle can create high frequency noise. These observations of excess high frequency noise are consistent with previous results. The SPL at 70 for both baseline and chevron nozzles for all secondary flow Mach numbers are shown in figure 10. SPL for different coaxial velocities have been vertically offset by an arbitrary amount for easy viewing. For M=0, significant reductions in mixing noise of approximately 3 dB due to the chevron nozzle are evident at the lower frequencies (1000 5000 Hz). There is a high frequency cross over in mixing noise above approximately 20,000 Hz. For M=0.28, low frequency noise reductions are slightly diminished to approximately 2.5 dB. A high frequency penalty can be seen above 20,000 Hz. For M=0.50, the low frequency reductions are approximately 1 dB. There is a high frequency penalty above 20000 Hz. For M=0.85, there are no low frequency reductions. The high frequency penalty seen above 20000 Hz is not as severe as for the lower Mach numbers. The variations to low and high frequency mixing noise are consistent with the different values of shear velocity across a chevron nozzle as discussed above. A 4 dB discrete tone can be seen with the baseline nozzle for M=0 at slightly above 4000 Hz. The chevron nozzle has reduced the level of screech to approximately 2dB. At M=0.28, there is a strong discreet tone (~13 dB above background level) associated with the baseline nozzle at slightly less than 4000 Hz. A discrete tone associated with the chevron nozzle is visible at the same frequency. However, its amplitude has been reduced to approximately 3dB. For M=0.50, there is a weak screech tone at approximately 3100 Hz which is largely eliminated by the chevron nozzle. There are no screech tones created for M=0.85. Screech tones are exacerbated by a reflecting surface, such as the trailing edge of a nozzle. Chevrons alter the trailing edge of a nozzle and do not provide reflecting surfaces that are normal to the flow. This change in geometry may be responsible for disrupting the feedback loop and as a result, altering the screech tones. For M=0, 0.28, and 0.50, the peak SPL for shock noise associated with the chevron nozzle is higher than the peak SPL for shock noise of the baseline nozzle. Moreover, the chevron nozzle is 0.6, 1.3, and 0.8 dB (OASPL) louder than the baseline nozzle for M=0, 0.28, and 0.50, respectively. For M=0.85, the peak SPL value of shock associated noise from the chevron nozzle is lower than that of the baseline nozzle. The chevron is also lower by 2.1 dB (OASPL). Shock noise is created by the interaction of turbulence in the jet plume with the quasi periodic shock cells that are created by an imperfectly expanded jet. Due to the increased mixing, a chevron will accelerate the decay of the potential core and the shock cell pattern. The rate at which this decay is increased will be proportional to the shear velocity across the chevron nozzle. A low shear will increase the rate a small degree. A high shear will increase the decay a large degree. Alternatively, chevrons will also increase the level of turbulence in regions directly downstream of the nozzle. A low shear will increase the turbulence slightly. A high shear will increase the turbulence significantly. The purpose of this work is to determine whether or not a favorable tradeoff between shock cells and turbulence is possible. As can be seen from the data, a favorable balance has been achieved for M=0.85. At lower secondary Mach numbers, the decrease in the shock cell pattern was not sufficient to offset the increase in turbulence. In addition, it can be seen for all secondary Mach numbers that the center frequency of the shock associated noise for the chevron nozzle is higher than that of the baseline nozzle. Use of the chevron nozzle significantly reduces the extent of the potential core and also the shock cell spacing. As can be seen from equation (1), any reduction in the shock cell spacing will increase the center frequency of shock associated noise, which is the effect that is seen in the far-field data. The shock cell spacing will be discussed in greater detail in a different section. It can also be seen for every secondary flow condition that chevrons reduce the spectral half width of shock associated noise. This is the same effect that was seen with a baseline nozzle when a secondary flow was added. For that case, reduced spatial jitter of the shock cell pattern and also the mean flow convection effect were responsible. Chevrons reduce the shear velocity across the jet plume. 7 American Institute of Aeronautics and Astronautics However, they also create excess turbulence immediately downstream of the nozzle which would increase jitter the of the most upstream shock cells. Tam16 has suggested that increased turbulence through the most upstream shock cells would increase will increase the jitter of all of the shock cells. For this reason, the mean flow convection effect may have more of an influence on the spectral half width than shock cell jitter. It was pointed out in the introduction that increasing secondary flow speed changes the velocity profile. Specifically, higher coaxial velocities reduce the shear velocity across the flow. This reduces the mixing between the two streams which increases the length of the potential core and also shock cell spacing. This reduces the peak/center frequency of the shock noise. In addition, the convection effect tends to decrease the spectral half width of the shock noise. These points were confirmed with experimental data. It was shown that for all cycle conditions tested, chevrons increased the center/peak frequency of the shock associated noise and also reduced the spectral half width. Chevrons increase mixing between adjacent streams. This reduces the length of the potential core and also the shock cell spacing. This would tend to increase the center/peak frequency of the shock noise, as was seen with the data. It was proposed earlier in this work that in order to reduce shock associated noise, there had to be a balance between reduced peak velocities and increased turbulence that results from the chevrons mixing action. Speaking generally, that balance is most easily achieved by varying two parameters, chevron penetration or shear velocity across the chevron nozzle. This work is limited to variations in shear velocity. At the highest shear velocity (secondary Mach number = 0), the chevron nozzle created approximately 1 dB more shock noise than the baseline nozzle. This would suggest that at this cycle condition, the reduction in shock cell strength due to the chevrons was offset by the increase in turbulence. At a slightly reduced shear velocity (secondary Mach number = 0.28), the shock noise associated with the chevron nozzle was approximately 2 dB higher compared to the baseline nozzle. As before, the higher shock noise of the chevron nozzle would suggest that the reduction in shock cell strength was again offset by the increase in turbulence. The result of another reduction in shear velocity (secondary Mach number of 0.50) was that the shock noise associated noise of the chevron nozzle exceeded that of the baseline nozzle by only 1 dB. At the lowest value of shear (secondary Mach number = 0.85) the shock noise associated with the chevron is now almost 3 dB below that of the baseline nozzle. The proper balance between the shock cell weakening and turbulence production has been achieved. However, it is curious that the difference between the shock noise levels would increase and then decrease with constantly decreasing shear velocity. Both the mixing of the jet and the production of turbulence are related to the shear velocity. However, their sensitivities to shear velocity may not be the same. Increasing the secondary flow from Mach 0 to 0.28 may have more of an impact on the production of turbulence than on the mixing of the jet. Increasing the secondary flow from Mach 0.28 to 0.50 may have more of an impact on mixing the jet than on the production of turbulence. The combination of these nonlinear effects would lead to the fluctuations in shock noise reduction that have been reported. However, there may be a contributing explanation that will be discussed in a later section. Discrete tones can be seen with both baseline and chevron data at the three highest shear velocity conditions (secondary Mach numbers = 0, 0.28, 0.50). Discrete, or screech, tones are a phenomenon associated with supersonic flow and are created by a feedback loop. In general, the chevrons were responsible for reducing the level of screech at a given cycle condition. In figure 9, the chevron reduced a 5 dB screech tone at 4000 Hz to less than 3 dB. In figure 10, the chevrons reduced a 13 dB screech tone at 4000 Hz down to less than 3 dB. In figure 11, chevrons reduced a 3dB screech tone at 3100 Hz down to 1 dB. However, in certain instance, the chevron nozzle actually increased the level of a screech tone. In figure 10, the chevrons increase a 3 dB screech tone at 4700 Hz to approximately 5dB. In figure 11, the chevrons increase a 3 dB screech tone at 4100 Hz to approximately 5dB. Finally, in that same figure, the chevrons create a 5dB screech tone at 4700 Hz where no screech tone had existed with the baseline nozzle. In general, discrete tones are exacerbated by a reflecting surface, such as the trailing edge of a nozzle. Chevrons alter the trailing edge of a nozzle and provide significantly less in the way of reflecting surfaces that are normal to the flow. However, while the surface normal to the flow may have decreased the total surface area of the trailing has increased significantly. These changes in geometry may be responsible for disrupting the feedback loop and altering the discrete tones as discussed. Earlier, it was suggested that there may be an additional explanations for the differences that were seen in shock cell noise between the baseline and chevron nozzles. For secondary Mach=0 (figure 10), an examination of the flow data showed that the primary stream NPR was actually 0.03% higher for the baseline configuration compared to the chevron configuration. Based on the discussion from the previous paragraph, that would indicate that the peak levels 8 American Institute of Aeronautics and Astronautics in the baselines shock noise should be even lower by as much as 0.5 dB. If the data were corrected to reflect that, then the shock noise of the chevron nozzle would be as much as 2 dB higher than the shock noise for the baseline nozzle. For secondary Mach=0.28 (figure 10), an examination of the flow data showed that the primary stream NPR was 0.09% lower for the baseline configuration compared to the chevron configuration. Loosely interpolating, that would indicate that the peak levels in the baselines shock noise could be approximately 1 dB higher. If the data were corrected to reflect that, then the chevron shock noise would be only 1.5 dB greater than the baseline shock noise. For secondary Mach number = 0.50 (figure 10), an examination of the flow data showed that the primary stream NPR was 0.14% higher for the baseline configuration compared to the chevron configuration. Loosely interpolating again, that would indicate that the shock cell noise from the baseline nozzle could actually be an additional 1 dB, or more, lower. Correcting the data to reflect that would mean that the chevons shock noise was approximately 2dB higher than the baseline shock noise. For secondary Mach number = 0.85 (figure 10), the primary stream NPR was 0.01% higher for the baseline configuration compared to the chevron configuration. Loosely extrapolating, that variation in NPR would suggest a correction of less than 0.2 dB. Given that the difference in shock cell noise between the two configurations is well over 2dB, that small amount will be ignored. This correction in shock noise with variations in NPR is interesting. However, there still does not appear to be a straightforward relationship between the differences in shock cell noise of the baseline versus chevron configurations and changing shear velocity across the primary nozzle. The maximum shear velocity (secondary Mach = 0) reveals that the peak shock noise from the chevron nozzle is 2 dB higher than the peak shock noise from the baseline nozzle. Reducing the shear velocity (secondary Mach = 0.28) reveals that the peak shock noise from the chevron nozzle is only 1.5 dB above that of the baseline nozzle. Further reduction in shear velocity (secondary Mach number = 0.50) reveals that the peak shock noise from the chevron nozzle is back to 2 dB above that of the baseline nozzle. The lowest shear velocity (secondary Mach number = 0.85) reveals that the peak shock noise from the chevron nozzle is now 2 dB below that of the baseline nozzle. Even with the uncertainties due to variation in NPR, the differences in measured shock noise are sufficient to unambiguously arrive at the following conclusions, for this particular chevron nozzle. For secondary flow Mach numbers from 0 to 0.50, chevrons increase the shock associated noise. For a secondary flow Mach number of 0.85, chevrons reduce shock associated noise. B. Shock Cell Spacing Measurements of the shock cell spacing for the baseline and chevron nozzles are shown in figure 11 and compared with references 12-14. The spacing is an average of the first five shock cells. Flight, or secondary flow Mach number, is on the abscissa while normalized shock cell spacing is shown on the ordinate. The shock cell spacing for M=0 has been estimated using equation (2) for each data set (except the chevron nozzle) and is also shown in the figure. These single data points use the same open symbol as their respective data set and are located at M=0. Norum & Shearin12 used a convergent nozzle discharging at a fully expanded Mach number = 1.8. These data demonstrate that shock cell spacing increases with coaxial flow Mach number. Equation 2 estimate for shock cell spacing in this case is approximately 6% low. Norum and Brown14 used a C-D nozzle with design Mach number of 1.5 discharging at a fully expanded Mach number of 1.8. These data also show that shock cell spacing increases with coaxial flow Mach number. The estimate for shock cell spacing is 5% high. Norum & Shearin13 used a convergent nozzle exhausting at a fully expanded Mach number of 1.166. An increased spacing with secondary flow speed can be seen for these data as well. The shock cell spacing estimate using equation (2) is 15% high. Baseline data from the current work also used a convergent nozzle exhausting at a fully expanded Mach number of 1.18. Because the conditions are so similar, the shock cell spacing is almost identical to Norum & Shearin13. These data show that shock cell spacing increase with secondary flow Mach number. The estimate for shock cell spacing is 22% high. For all five sets of data, shock cell spacing was shown to increase with secondary Mach number. In addition, it would appear that equation (2) is more accurate for flows with higher, fully expanded velocity. C. Near-Field Acoustic Near-field acoustic results at 1250 Hz for the baseline nozzle and all four coaxial flow conditions are shown in figures 12a-e. Reference to figures 8 and 9 indicate that this frequency has no component of shock noise, it is exclusively mixing noise. The plot for secondary Mach = 0 (figure 12a) is dominated by a large region of noise between 4-18 Deq. With a secondary flow of Mach 0.28 (figure 12b), the noise region is significantly reduced. For secondary flow Mach 0.50 (figure 12c), the noise is further reduced and is now a minimum. The noise increases with a secondary flow of Mach 0.85 (figure 12d). This trend in mixing noise is the same as what was seen with the far-field results for the baseline nozzle. Increasing the secondary flow from Mach 0 to 0.28 reduced the strength of the inner shear layer which created less jet noise. The outer shear layer did not make enough noise to be heard 9 American Institute of Aeronautics and Astronautics compared to the inner shear layer. Increasing the speed of the secondary flow to Mach 0.50 further reduces the strength of the inner shear. The outer shear layer is not yet making significant noise compared to the inner shear layer. Increasing the secondary flow to Mach 0.85 further weakens the inner shear layer. However, the outer shear layer now dominates. This is shown more qualitatively in figure 12e. The data points along and directly adjacent to the jet plume have been extracted. This way, SPL at different axial displacements for the four cycle conditions can be compared. For secondary Mach = 0, peak noise occurs at X/Deq~11. Increasing secondary flow to Mach 0.28 reduces the strength of the inner shear layer as can be seen. Further, it retards the growth of the shear layer and it peaks farther downstream at X/Deq~13. Increasing secondary flow to Mach 0.50 further reduces the growth of the shear layer and peak noise is now seen at X/Deq~16. With the secondary flow at Mach 0.85, the outer shear layer dominates. The secondary flow exit plane is upstream of the primary exit plane and peak noise is centered at x/Deq~6. Near-field acoustic results for the baseline nozzle at 8000 Hz and all cycle points are shown in figures 13a-e. Figure 10 demonstrates that this frequency is approximately the center/peak frequency for shock noise when the secondary flow is Mach 0, 0.28, and 0.50. Although the center/peak frequency for secondary flow Mach = 0.85 is lower at approximately 6500 Hz, shock noise still dominates at 8000 Hz and it will be used. Figure 13a shows the near-field mapping for secondary flow Mach =0. There is a dominant region of noise that extends from 2-8 Deq. Increasing the secondary flow to Mach 0.28, figure 13b, reduces the level of noise and also shifts the generation region downstream. It now extends from approximately 5-9 Deq. The region for secondary Mach = 0.50, figure 13c, reduces the noise to a minimum. When secondary flow Mach = 0.85, the noise region is now much more significant and has shifted downstream to approximately 8-18 Deq. These trends can also be seen quantitatively in figure 13e. It is more apparent with this figure that shock noise is centered farther downstream and is at lower SPL when secondary flow Mach number increases from 0 to 0.50. Shock noise is centered even farther downstream but at a much higher level when secondary flow Mach number = 0.85. These trends in the level of shock associated noise as secondary flow increases are consistent with far-field results. Moreover, it can be seen that shock noise extends downstream with increasing secondary flow velocity. This would suggest that the secondary flow is stretching the shock cell pattern farther downstream, which is consistent with static pressure measurements. Near-field acoustic results for the chevron nozzle at 1250 Hz and all coaxial flow conditions are shown in figure 14a-e. Reference to figures 7 and 8 demonstrate that only mixing noise is created at this frequency. Results for secondary Mach = 0 are shown in figure 14a. The dominant region of noise generation extends from 8-14 Deq. The noise is reduced as the secondary flow Mach increase to 0.28. It is a minimum for secondary Mach = 0.50. The region increases in extent and also magnitude at secondary Mach = 0.85. The trends are more clearly seen in figure 14e, which shows SPL values that are adjacent to the jet plume for all four cycle conditions. The mixing noise drops to a minimum as secondary flow Mach number increases from 0 to 0.50. The levels then climb to a maximum for secondary flow Mach = 0.85. The trends seen with the near-field plots reflect the trends that were seen with the farfield plots. The near-field results for the chevron nozzle at higher frequency and all coaxial flow conditions are shown in figure 15a-e. The purpose of this section is to compare the shock noise levels for the chevron nozzle at all four cycle conditions. Reference to figure 8 would suggest that 8000 Hz is reasonably close to the center frequency for secondary flow Mach numbers of 0, 0.28, 0.50. However, the center frequency for secondary Mach = 0.85 is approximately 2000 Hz lower. In addition, the SPL at 8000 Hz is approximately 5 dB below the peak. Therefore, the data for 6300 Hz will be used instead. The near-field results for secondary flow Mach = 0 are shown in figure 15a. A dominant region of noise can be seen at X/Deq ~ 1-6. Figure 15b shows that the magnitude of the generation region is slightly lower and it shifts downstream to approximately 3-7 Deq when secondary flow Mach = 0.28. Figure 15c demonstrates that the region drops to a minimum for secondary Mach = 0.50. With secondary Mach = 0.85, the region is now much more dominant and has shifted farther downstream. Figure 15e shows the data in a more quantitative form. It can be seen that the higher SPL levels occur with secondary Mach = 0 and fall to a minimum when secondary Mach = 0.50. SPL levels rise again with secondary Mach = 0.85. The trends in noise level are similar to far-field results. In addition, the downstream extension of shock noise is consistent with static pressure measurements of the jet plume. Near-field acoustic results for both baseline and chevron nozzle and all cycle points are discussed next. The frequencies that will be shown are 10,000, 10,000, 8,000, and 5,000 Hz. They correspond to secondary flow Mach numbers of 0, 0.28, 0.50, and 0.85 respectively. Reference to figure 10 demonstrate that all of these frequencies are 10 American Institute of Aeronautics and Astronautics dominated by shock noise at their respective Mach numbers. Results for Mach=0 are shown in figures 16a & b. The near-field mapping for the baseline nozzle is dominated by a high intensity noise region that is located between 4 and 9 Deq downstream of the primary nozzle. There is a similar region of high intensity noise in the chevron mapping. However, it starts farther upstream at ~0.5 Deq but only extends to 7 Deq. More importantly, it appears to be more intense. The SPL values for the inner most row of the mapping grid are shown in figure 16c. Shock noise levels from the chevron nozzle exceed that of the baseline nozzle by as much as 5dB from the primary nozzle exit to approximately 6 Deq. Farther downstream, all levels are much lower but the baseline nozzle now exceeds the chevron nozzle by as much as 5dB. The near-field mappings at Mach 0.28 for baseline and chevron nozzles are shown in figures 17a & b. This mapping has a region of high intensity noise located between 4 and 11 Deq downstream of the primary nozzle. The noise region behind the chevron nozzle is higher in intensity and extends from the primary nozzle exit plane to approximately 11 Deq. Figure 17c compares the inner most row of SPL for the two nozzles. Shock noise levels behind the chevron nozzle exceed the baseline nozzle by as much as 4 dB until 8 Deq. The near-field mappings at Mach 0.50 for the baseline and chevron nozzles are shown in figures 18a & b. The high intensity region of noise behind the baseline nozzle extends between 4 and 13 Deq. The noise region behind the chevron nozzle extends from the primary nozzle exit plane to only 11 Deq. Shock noise from the chevron nozzle exceeds that of the baseline nozzle by as much as 4 dB. A number of trends are apparent from these tests results. In every case, shock noise behind the chevron nozzle started at least 3 Deq upstream of the start of shock noise behind the baseline nozzle. Shock noise is created by the interaction of turbulence and shock cells. Chevrons create excess turbulence immediately downstream of the nozzle, particularly at these relatively high shear conditions. This increased turbulence creates shock noise at more upstream locations compared to the baseline nozzle. Additionally, shock noise behind the baseline nozzle extends at least 1.5 Deq farther downstream compared to shock noise behind the chevron nozzle. When chevrons increase mixing, they reduce the length of the potential core and also the number of shock cells. This limits the extent of shock associated noise creation, which is demonstrated by the nearfield plots for these three secondary flow Mach numbers. Most importantly, the region of noise behind the chevron nozzle is higher in intensity compared to the region behind the baseline nozzle for all three secondary flow Mach numbers. The near-field mappings of baseline and chevron nozzles at Mach number 0.85 are shown in figures 19a & b. A region of intense noise behind the baseline nozzle extends from 7 to 15 Deq. Alternatively, the region of high intensity noise behind the chevron nozzle does not start until 8 Deq and is no longer present by 12 Deq. As seen on figure 19c, shock noise behind the baseline nozzles exceeds that behind the chevron nozzle by about 1 dB until X/Deq~9. Downstream of that, the baseline nozzle shock noise exceeds that from the chevron nozzle by as much as 5dB. The chevron produces less shock noise at all locations compared to the baseline nozzle at this cycle point. As with previous cycle conditions, the noise region behind the baseline nozzle extends farther than the region behind the chevron. This is caused by the reduced length of the potential core and shock cells as discussed previously. However, unlike the previous cycle conditions, the region of noise behind the chevron nozzle starts 1 Deq farther downstream compared to the baseline nozzle. Most importantly, the region of noise behind the chevron nozzle is lower in intensity compared to the region behind the baseline nozzle. The near-field mappings demonstrate that shock noise behind a chevron nozzle is more intense than the shock noise behind the baseline nozzle for the three, lower Mach numbers. They also demonstrate that the shock noise behind the chevron nozzle is less intense compared to the baseline nozzle at a simulated flight of Mach 0.85. This is consistent with far-field measurements. IV. Summary and Conclusions Far-field acoustic tests were performed to study the influence of chevron nozzles on the shock-associated noise of an underexpanded jet. The primary (core) stream of a coaxial test facility was set to a constant NPR of 2.36. The secondary (coaxial or fan) stream was varied between NPRs of 1 and 1.604 to study the effect of secondaryl flow velocity on noise from the coaxial jet. It was shown that in static and low secondary flow conditions, the shock cell noise increased when the chevron nozzle was used. Alternatively, at the highest secondary NPR, the shock associated noise was decreased. These results suggest that at the lower fan flow velocity, the higher level of turbulence created by the chevron nozzle offsets any reductions in shock cell strength and creates more shock associated noise. However, at high secondary flows, the reduction in shock cell strength due to the chevron has offset the increased level of turbulence and lowered the level of shock associated noise. In general, chevrons were shown to reduce discrete screech tones. It was suggested that the altered trailing edge of chevron nozzles and the reduced coherence of flow structures generated there disrupts the feedback loop, which is responsible for creating these discrete tones. 11 American Institute of Aeronautics and Astronautics Near-field acoustic tests demonstrated that the downstream extent of shock associated noise was always reduced when a chevron nozzle was used. This occurs because the chevron increases the mixing of the primary flow into the secondary flow. This reduces the length of the potential core and also eliminates the most downstream shock cells. Variations in the upstream extent of shock noise were also documented. It was suggested that increased turbulence created by the chevron nozzle interacted with the existing shock cells to enhance shock noise. Near-field mappings at secondary Mach numbers = 0, 0.28, and 0.50 demonstrated that shock noise associated with the chevron nozzle was more intense than shock noise associated with the baseline nozzle. Near-field mappings at secondary Mach number = 0.85 demonstrated that shock noise associated with the chevron nozzle was less intense than shock noise associated with the baseline nozzle. Both of these observations are in complete agreement with the far-field results. Static pressure measurements from a Pitot probe were presented and discussed. It was shown that the shock cell spacing of the baseline nozzle compared favorably with previously published data at a similar cycle points. It was also shown that increased secondary flow, approximating effects of flight velocity, increased the shock cell spacing. This leads directly to the reduced center/peak frequency that was seen in the far-field measurements and is in accordance with equation (1). It was shown for all cycle points that the shock cell spacing behind a chevron nozzle was less than that of the baseline nozzle. This leads directly to an increased center/peak frequency that was seen in the far-field measurements and is also in accordance with equation (1). The results from these tests suggest that chevrons can be used to reduce shock associated noise for high speed secondary flows. These conditions can be used to deduce effects of chevron during cruise conditions at a similar Mach number. However, only one chevron penetration was surveyed. In his work, Callender et al3 suggested that due to the number of different cycles and shear velocities on modern turbofan engines, the optimum chevron penetration probably varies from engine to engine. As was demonstrated, shock noise is particularly sensitive to variations in NPR, and is likely to be just as sensitive to chevron penetration. Therefore, optimum chevron penetration for broadband shock noise reduction must probably be tailored to particular engines and engine cycles. This was also suggested by Long24. References 1 Janardan, Hoff, Barter, Martens, Gliebe, Mengle, Dalton, AST Critical Propulsion and Noise Reduction Technologies for Future Commercial Subsonic Engines, Separate-Flow Exhaust System Noise Reduction Concept Evaluation, Final Report, NAS3-27720, Area of Interest 14.3, 1998 2 Saiyed, N., Mikkelsen, K., Bridges, J., Acoustics and Thrust of Separate-Flow Exhaust Nozzles with Mixing Devices for High Bypass-Ratio Engines, AIAA 2000-1961, 2000 3 Callender, W.B., Gutmark, E., Martens, S., A Far-Field Acoustic Investigation of Chevron Nozzle Mechanisms and Trends, AIAA 2003-1058, 2003 4 Callender, W.B., Gutmark, E., Martens, S., A Near-Field Investigation of Chevron Nozzle Mechanisms, AIAA 20033210, 2003 5 Callender, W.B., Gutmark, E., Martens, S., A PIV Flow Field Investigation of Chevron Nozzle Mechanisms, AIAA 20040191, 2004 6 Rask, O.H. Gutmark, E., Martens, S., Acoustic Investigation of a High Bypass Ratio Separate Flow Exhaust System with Chevrons, AIAA 2004-0009, 2004 7 Andersen, J. D., Modern Compressible Flow with Historical Perspective, 2nd Edition,McGraw-Hill, New York, 1990 8 Harper-Bourne, M., Fisher, M., J., The Noise from Shockwaves in Supersonic Jets, Proceedings (No. 131) of the AGARD Conference on Noise Mechanisms, Brussels, Belgium, 1973 9 Tanna, H., K., An Experimental Study of Jet Noise, Part II, Shock Associated Noise, Journal of Sound and Vibration, Vol. 50, 1977, pp 429-444 10 Tam, C.K.W., Tanna, H., K., Shock Associated Noise of supersonic jets from convergent-divergent nozzles, Journal of Sound and Vibration, Vol. 81, 1982, pp 337-358, 11 Norum T.,D., Seiner, J.,M., Measurements of mean static pressure and far field acoustics of shock-containing supersonic jets, TM 84521, 1982 12 Norum T., D., Shearin, J., G., Effects of Simulated Flight on the Structure and Noise of an Under expanded Jet, NASA TP-2308, 1984 13 Norum T., D., Shearin, J., G., Shock Structure and Noise of Supersonic Jets in Simulated Flight to Mach 0.4, NASA TP2785, 1988 14 Norum, T., D., Brown, M., C., Simulated High Speed Flight Effects on Supersonic Jet Noise, AIAA 93-4388, 1993 15 Brown, W.H., Ahuja, K.,K., Tam, C.,K.,W., High Speed Flight Effects on Shock Associated Noise, AIAA 86-1944, 1986 16 Tam, C.K.W., Supersonic Jet Noise, Annual Review of Fluid Mechanics, Vol. 27, 1995, pp 17-43, 17 Tam, C.,K.,W., The Near and Far Acoustic Fields of Broadband Shock Associated noise, AIAA 86-1943, 1986 12 American Institute of Aeronautics and Astronautics 18 Tam, C.,K.,W., Seiner, J.M., and Yu, J.C., Proposed Relationship Between Broadband Shock Associated Noise and Screech Tones, Journal of Sound and Vibration, Vol. 110, 1986, pp 309-321, 19 Tam, C.K., Stochastic Model Theory of Broadband Shock Associated Noise from Supersonic Jets, Journal of Sound and Vibration, Vol. 116, 1987, 265-302, 20 Tam, C.K., Broadband Shock-Associated Noise of Moderately Imperfectly Expanded Supersonic Jets, Journal of Sound and Vibration, Vol. 140, 1990, pp 55-71 21 Tam, C.K.W., Forward Flight Effects on Broadband Shock Associated Noise of Supersonic Jets, AIAA 89-1088, 1989 22 Tam, C.K.W., Broadband Shock-Associated Noise from Supersonic Jets in Flight , Journal of Sound and Vibration, Vol. 151, 1991, pp 131-147, 23 Long, D., F., The Structure of Shock Cell Noise from Supersonic Jets, AIAA 2005-2840, 2005 24 Long, D., F., Effect of Nozzle Geometry on Turbofan Shock Cell Noise at Cruise, AIAA 2005-0998, 2005 25 Callender, W.B., Gutmark, E., Dimicco, R., The Design and Validation of a Coaxial Nozzle Acoustic Test Facility, AIAA 2002-0369, 2002 13 American Institute of Aeronautics and Astronautics nozzl p e Slip line Expansion fan Shock wave Flow direction pe Figure 1. Two cells of a quasi periodic shock cell pattern. 100 5 dB 95 SPL (dB re 20 Pa) 150 90 90 85 80 75 100 1000 Frequency (Hz) 10000 100000 Figure 2. Sensitivity comparison, Primary NPR=2.36, Secondary NPR = 1.00 14 American Institute of Aeronautics and Astronautics 105 5 dB 100 SPL (dB re 20 Pa) 95 150 90 90 85 80 75 100 1000 Frequency (Hz) 10000 100000 Figure 3. Sensitivity comparison, Primary NPR = 2.36, Secondary NPR = 1.604. 125 150 115 10 dB 140 130 SPL (dB re 20Pa) 105 120 110 95 100 85 90 70 75 100 1000 Frequency (Hz) 10000 100000 Figure 4. SPL of baseline nozzle for secondary flow of M = 0 and primary flow at NPR = 2.36. The arrows indicate the frequency predicted from Equation (1), with uc=73% of fully expanded velocity. Individual spectra are displaced vertically for easy viewing. 15 American Institute of Aeronautics and Astronautics 105 5 dB 100 SPL (dB re 20 Pa) 95 90 85 80 Secondary Mach = 0.85 Secondary Mach = 0.50 Secondary Mach = 0.28 Secondary Mach = 0.00 75 100 1000 Frequency (Hz) 10000 100000 Figure 5. SPL of baseline nozzle at 150. Core nozzle at NPR=2.36 with variable secondary flow. 95 5 dB 90 SPL (dB re 20Pa) 85 80 75 70 Secondary Mach = 0.85 Secondary Mach = 0.50 Secondary Mach = 0.28 Secondary Mach = 0.00 65 1000 Frequency (Hz) 10000 100000 Figure 6. SPL of baseline nozzle at 90. Core nozzle at NPR=2.36 with variable secondary flow. The arrows indicate the frequency predicted from Equation (1), with uc=73% of fully expanded velocity plus 30% of flight velocity. 16 American Institute of Aeronautics and Astronautics 105 5 dB 100 SPL (dB re 20 Pa) 95 90 85 80 Secondary Mach = 0.85 Secondary Mach = 0.50 Secondary Mach = 0.28 Secondary Mach = 0.00 75 100 1000 Frequency (Hz) 10000 100000 Figure 7. SPL of chevron nozzle at 150. Core nozzle at NPR=2.36 with variable secondary flow. 95 5 dB 90 SPL (dB re 20 Pa) 85 80 75 70 Secondary Mach = 0.85 Secondary Mach = 0.50 Secondary Mach = 0.28 Secondary Mach = 0.00 1000 10000 100000 65 100 Frequency (Hz) Figure 8. SPL of chevron nozzle at 90. Core nozzle at NPR=2.36 with variable secondary flow. 17 American Institute of Aeronautics and Astronautics 105 M = 0.85 95 10 dB M = 0.50 M = 0.28 M = 0.00 baseline chevron SPL (dB re 20Pa) 85 75 65 55 45 100 1000 Frequency (Hz) 10000 100000 Figure 9 SPL of baseline and chevron nozzles at 150 for all secondary Mach numbers. Primary NPR=2.36. 95 M = 0.85 85 10 dB baseline chevron SPL (dB re 20Pa) 75 M = 0.50 65 M = 0.28 55 M = 0.00 45 35 1000 10000 Frequency (Hz) 100000 Figure 10. SPL of baseline and chevron nozzles at 70 for all secondary Mach numbers. Primary NPR=2.36. 18 American Institute of Aeronautics and Astronautics 2.5 Normalized Shock Cell Spacing 2.0 Present Data: Baseline Present Data: Chevron 1.5 Estimated shock cell spacing using equation (2) Norum and Brown, Ref 14 Norum & Shearin, Ref 13 Norum & Shearin, Ref 12 1.0 0.5 0.0 0.0 0.2 0.4 0.6 Flight Mach Number 0.8 1.0 Figure 11. Shock cell spacing for different secondary flow/flight Mach number. 19 American Institute of Aeronautics and Astronautics (a) (c) (b) (d) 130 125 SPL (re 20 Pa) (e) 120 115 110 Secondary Mach = 0.85 Secondary Mach = 0.28 Secondary Mach = 0.50 Secondary Mach = 0.00 105 -2 0 2 4 6 8 10 Xeq 12 14 16 18 20 22 Figure 12. Near-field plots for baseline nozzle at 1250 Hz a) Secondary Mach number = 0 b) Secondary Mach number = 0.28 c) Secondary mach number = 0.50 d) Secondary Mach number = 0.85 e) SPL along inner contour 20 American Institute of Aeronautics and Astronautics (a) (c) (b) (d) 130 (e) SPL (re 20 Pa) 125 120 115 110 Secondary Mach = 0.85 Secondary Mach = 0.28 Secondary Mach = 0.50 Secondary Mach = 0.00 105 -2 0 2 4 6 8 10 Xeq 12 14 16 18 20 22 Figure 13. Near-field plots for baseline nozzle at 8000 Hz a) Secondary Mach number = 0 b) Secondary Mach number = 0.28 c) Secondary mach number = 0.50 d) Secondary Mach number = 0.85 e) SPL along inner contour 21 American Institute of Aeronautics and Astronautics (a) (c) (b) (d) 130 (e) 125 SPL (re 20 Pa) 120 115 110 Secondary Mach = 0.85 Secondary Mach = 0.28 Secondary Mach = 0.50 Secondary Mach = 0.00 105 -2 0 2 4 6 8 10 Xeq 12 14 16 18 20 22 Figure 14. Near-field plots for chevron nozzle at 1250 Hz a) Secondary Mach number = 0 b) Secondary Mach number = 0.28 c) Secondary mach number = 0.50 d) Secondary Mach number = 0.85 e) SPL along inner contour 22 American Institute of Aeronautics and Astronautics (a) (c) (b) (d) 130 (e) 125 SPL (re 20 Pa) 120 115 110 Secondary Mach = 0.85 Secondary Mach = 0.28 Secondary Mach = 0.50 Secondary Mach = 0.00 105 -2 0 2 4 6 8 10 Xeq 12 14 16 18 20 22 Figure 15. Near-field plots for chevron nozzle a) Secondary Mach number = 0, 8000 Hz b) Secondary Mach number = 0.28, 8000 Hz c) Top Secondary mach number = 0.50, 8000 Hz d) Secondary Mach number = 0.85, 6300 Hz e) SPL along inner contour 23 American Institute of Aeronautics and Astronautics 130 Baseline Chevron 125 SPL (dB re 20 Pa) 120 115 110 -2 0 2 4 6 8 10 Xeq 12 14 16 18 20 22 Figure 16. Near-field plots at 10000 Hz for baseline and chevron nozzle, secondary M=0. 24 American Institute of Aeronautics and Astronautics 130 Baseline Chevron 125 SPL (dB re 20 Pa) 120 115 110 -2 0 2 4 6 8 10 Xeq 12 14 16 18 20 22 Figure 17 Near-field plots at 10000 Hz for baseline and chevron nozzle, secondary Mach=0.28 25 American Institute of Aeronautics and Astronautics 125 Baseline Chevron 120 SPL (dB re 20 Pa) 115 110 105 -2 0 2 4 6 8 10 Xeq 12 14 16 18 20 22 Figure 18. Near-field plots at 8000 Hz for baseline and chevron nozzle, secondary Mach=0.50. 26 American Institute of Aeronautics and Astronautics 130 Baseline Chevron 125 SPL (dB re 20 Pa) 120 115 110 -2 0 2 4 6 8 10 Xeq 12 14 16 18 20 22 Figure 19. Near-field plots at 5000 Hz for baseline and chevron nozzle, secondary Mach=0.85. 27 American Institute of Aeronautics and Astronautics
MOST POPULAR MATERIALS FROM IS
MOST POPULAR MATERIALS FROM Cincinnati