1、NA-04-3-1 (RP-I 026) The Acoustic Properties of Common HVAC Plena Emanuel Mouratidis, P.Eng. Member ASHRAE ABSTRACT The HVAC plenum is commonly applied as a means to reduce problematic low-frequency noise produced by mechan- ical equipment in ducted systems. The predicted acoustical performance has
2、been derived from an earlier investigation Fells 1958). is early work did not focus on lowfrequency performance, the efects of wall linings, or typical geometric variables. Through the ASHRAE-sponsored research project RP-1026, conducted by Vibro-Acoustics, rigorous testing enabled the identlJicatio
3、n of comprehensive attenuation prop- erties for common HVAC plena. A two-stage prediction approach resulted, which applies an improved version of the Wells equation for mid- to high-fequency attenuation (5000 Hz upper frequency limit), quantified as a transmission loss (TL), and a surface area and w
4、all characteristic model for the critical low-fequency TL (50 Hz lower frequency limit). This method has improved the prediction accuracy of flow-through type HVACplena with a volume range of 20-1100 f (0.5-30 m3). In addition, this work has identijed acoustical effects for seven different wall-lini
5、ng systems and typical geometric variables, such as opening ofsets, opening ratio variables, and multiple openings. INTRODUCTION The acoustical plenum is a device used to attenuate sound, usually low frequency, within a ducted HVAC system. It is defined as a type of passive sound attenuator that pri
6、marily achieves its performance through a sudden area change. The resulting expansion effect creates high levels of reflections due to the acoustic impedance transfer, which is the ratio of pressure amplitude to the resulting particle velocity at a plenums entrance and termination. In typical HVAC s
7、ystems, John Becker the broadband attenuation is further enhanced by the applica- tion of absorptive wall lining materials. Plenum acoustic performance is characterized in terms of transmission loss (TL). The most common prediction tool is the equation found in the ASHRAE Handbook-HVAC Appli- cation
8、s (ASHRAE 1999), where Plenum TL is as follows: TL = -1O*log so, *(cose.-) (1) 4Kr2 Sau TL = A, (la) where ml, = s= r= a, = Ka = A, = e= area of outlet section of plenum, ft2 (m2) total inside surface area of plenum less the inlet and outlet areas, It2 (m2) distance between centers of inlet and outl
9、et sections, fi (4 directivity factor 2 (for an inlet opening at the approximate center of a wall) 4 (for an inlet opening adjacent to one wall or at a bi- hedral comer) 8 (for an inlet opening adjacent to two walls or at a tri- hedral comer) average absorption coefficient of plenum lining offset an
10、gle at the exit, relative to the inlet opening attenuation coefficient attenuation according to ASHRAE 1999 equation Emanuel Mouratidis and John Becker are with Vibro-Acoustics, Toronto, Ontario, Canada. 02004 ASHRAE. 597 Equation 1 is a derivation from the work by Wells (1958), which is a variation
11、 of the common room constant equation. It characterizes the total energy density at a point within the chamber as the sum of the direct and reverberant energy densi- ties. Wells work was based on the dynamic scaling of data from small test plena (no plenum dimension greater than 3 ft O.9m) and a 375
12、 Hz lower frequency limit. Inlet area, multi- ple outlet openings, specific location of outlet openings (e.g., side vs. end), and the presence of airflow were not considered as variables affecting the plenums performance characteris- tics. Furthermore, this early work did not identify the low- frequ
13、ency characteristics when in the presence of plane wave propagation. Since the fundamental property of a plane wave is that of constant pressure and displacement amplitude perpendicular to the direction of travel, the resulting three- dimensional interaction with a plenum chamber may be highly depen
14、dent on both the inlet duct and plenum dimensions that support these unique sound waves. In this early work, Wells described an anticipated error of IT10 dB for the frequencies below the range of cross-mode or multi-mode propagation, as determined by the cutoff frequency Xand Y = width and height of
15、 the inlet duct or opening, respectively, ft (m); and CO = speed of sound, Ws (ds). cut-off frequency is determined by For the (1,O) fundamental mode using square duct, the f =Co 2d (4) where d = duct dimension, width or height, ft (m). In this project, the inlet duct sizes selected (as noted above)
16、 corresponded with approximate cut-off frequencies of 570 Hz, 285 Hz, and 140 Hz, respectively. This represents cut- off frequencies for a wide range of inlet duct sizes commonly found in HVAC duct systems. To simpli9 the regression analysis, a power series was selected as the method to correlate th
17、e dimensionless Kfand the corresponding test data (dB) within the dual CfIfc and ffc) frequency regions. Figure 5 displays the distribution of A, versus Kf, which produced very good corresponding coe- ficients of determination (R2 = 0.78), above the critical cut-off region. As shown in Figure 6, the
18、 scatter among the low- frequency A, data versus Kfwas very large. This resulted in a very weak correlation coefficient (R2 = 0.40). As the 1999 ASHRAE Handbook-Applications chapter suggests, there is a significant dependence on cutoff frequency when applying Equation 1. ASHRAE Transactions: Symposi
19、a 599 waon&mw-,y (IWI I s i, m single metal wall 1 acoustical hing ,“, single IyeE_B metal wall 80 1 O0 125 ueL. single metal wall 2 acoustical lining 0.01 0.04 0.05 0.05 0.30 0.60 0.05 0.29 0.01 0.04 0.05 0.05 0.50 0.50 0.05 0.21 0.01 0.04 0.26 0.38 0.50 0.50 0.13 0.05 IyeE4 double metal walls uoy
20、IYeLE double metal walls double metal walls 4 acoustical blanket fill 8 acoustical blanket fill 4 no-media type 4 acoustical blanket fiil IyeLp double metal walls perf facing perf facing micro-perf facing solid facing 2500 3200 4000 5000 Figure 2 Typical wall construction details (section view). 0.0
21、2 0.05 0.94 0.97 1 .o0 1 .o0 0.21 0.05 0.03 0.07 0.92 0.95 0.94 0.96 0.18 0.05 0.03 0.07 0.88 0.91 0.93 0.93 0.13 0.05 0.03 0.07 0.81 0.83 0.86 0.86 0.05 0.05 Table 1. Random Incidence Wall Absorption Coefficients, a Metal Type D Double Wall w/ 4 in. Liner (perf/solid) 0.30 0.45 Type E Double Wall w
22、l8 in. Liner (perfholid) 0.35 0.45 TY Pe F Double Wall wl 4 in. Chamber (no liner) 0.05 0.08 Type G Double Wall wt 4 in. Liner (solid/solid) 0.26 0.48 600 ASHRAE Transactions: Symposia Table 2. Acceptable Error Tolerances, oiso SO 96141 %o I 1 113 Oct. Band Center Freq. I 50-160 I 200-315 I 400-5k I
23、 6.3k- 1Ok 3 2 1.5 2.5 i TYPICAL SOUND INTENSITY PLENUM OUTLET BLOCKAGE CONTROL VOLUME 71 USED TO DETERMINE Lp(bbckiigeZ 30 25 * *ATL k-d Min 10 X duct dia Figure 3 Test setup. Figure 4 ATL and A, for a 4 x 6 x 1 Oft Type-Cplenum. 45 40 e. 35 4 30 1 25 i: 10 5 O 00 O1 O2 03 04 05 OB 07 O8 O9 10 Awen
24、ustion CoatRoient U,) Figure 5 ATL versus Kf ( f, only) for all in-line Figure 6 ATL versus Kf (50f$Volume 1 (dB/f?) (dB/ft2) TypeB TypeC TypeD TYpeE TYPeF TYPeG 0.14 0.03 1 1 O 1 O O Wall Effect, We (dB ADD) 63 80 1 O0 125 160 0.10 0.03 1 2 3 7 1 3 0.11 0.03 2 2 3 9 2 7 0.23 0.03 2 2 4 12 1 6 0.24
25、0.04 2 3 6 12 1 4 0.20 0.04 3 4 11 11 O 2 (not applicable) 1 I - I I 20 i -15 -20 50 63 Bo rm i25 ia, 200 250 315 400 Mo Frequency (Hz) Figure 7 Lowfrequency wall effect for all Type-B configurations (AA = ATL Type-B - AT= Type-Pv). combined with the regression coefficients from Ap creates a hybrid
26、model to represent the broadband attenuation. Exclud- ing the potential effects of offset orientations, such as a cos0 term #l, the prediction of plenum attenuation is as follows. For 50 Hz Iff, and always increased with increasing 0. At f If, the scatter in the data was found to be very high and th
27、e magnitude of the change itself extremely low. This was especially true within the 63 Hz octave, where approximately no change in TL was found for the wide range of 0 studied. In this report, the elbow effect corresponds to the magni- tude change in attenuation due to a plenum configuration change
28、from an end-inletend-outlet with in-line openings to an end-inlethide-outlet openings (see Figure 1). The variability in the elbow effect data was very high in comparison with the other -5 J 1 0.5 0.6 0.7 0.8 0.9 1 .o 1.1 COSO Figure 11 Typical high frequency offset opening efects-dB change in ATL v
29、ersus COSO. geometric variables studied. The elbow effect appeared to be dependent onf, with a 2-3 dB increase in attenuation abovef,. For theflf, range, the scatter among the test data was signifi- cant. The elbow effect increased the attenuation in the 50- 160 Hz range and reduced the attenuation
30、in the mid-frequency range. The majority of in-line plenum configurations used in this study, as highlighted earlier in this paper, displayed an atten- uation peak or apparent tuning within various mid-frequency bands. As shown in Table 5, although the elbow effect created a 2-6 dB broadband increas
31、e in attenuation, independent of plenum size and construction type, the U3-0ctave bands within this potential tuning frequency range (200 to 500 Hz) were generally reduced. The elbow effect, as produced by an end- inletside-outlet configuration, may require further study using finer dimensional vari
32、ations. This may reliably identifj the apparent elimination of plenum tuning. The limited multi-outlet plenum configurations in this work produced negligible change in attenuation (k1.5 dB), as compared to a corresponding single outlet in-line or elbow plenum configuration. The analysis did not iden
33、ti5 a sound power division from one outlet to two or three, as in a 3 dB or 5 dB broad-band reduction, respectively. Therefore, once a critical path has been established in a plenum system, either in an in-line or elbow configuration, the presence of multiple openings may be ignored in the TL predic
34、tion. ASHRAE Transactions: Symposia 603 Table 4. Offset Angle Effects on TL for an End-Outlet Plenum Offset Angle (0) O 15.0 22.5 30.0 37.5 45.0 I 50 Io O O O O O 63 80 O O O O O O O O -1 -3 -4 -6 O 1 O -2 -3 -6 1 O -2 -4 -6 O O -1 -2 -3 -4 400 O 2 4 6 9 13 500 O 1 3 6 10 15 - O -1 -2 -3 -5 1 2 3 5
35、7 O 4 6 8 10 14 Offset Angle (0) O 15.0 22.5 30.0 37.5 45.0 I = 630 (not applicable) Offset Angle Effects on TL , .f,: I Freq. I C. = 160 (not applicable) 200 I 250 I 315 400 I 500 I 630 800 I 1000 I tli 2000 4000 I 5000 I O O O o O O O O O O O O O O O 1 2 1 1 O 1 1 1 O O 1 1 O O O 4 4 2 2 1 2 2 2 2
36、 1 2 2 2 2 3 9 8 3 3 2 3 2 4 4 1 4 3 4 5 6 14 13 4 4 4 5 3 6 6 2 7 5 6 8 10 20 19 5 6 5 7 3 9 9 3 10 8 9 12 15 604 ASHRAE Transactions: Symposia - 9 1 Freq. 50 63 80 1 O0 125 160 200 250 O 1 50 63 80 100 125 1M) 200 250 315 400 500 630 800 1000 1300 1600 2oM) 2500 3200 4000 5000 Frequency (Hz) fc =
37、630 2 (not applicable) I 630 I3 I I 1250 12 I I 5000 I 1 I The reduction in attenuation for an opening area ratio 1 .O was found to be 1 .O dB per the magnitude of the A-Ratio, where A-Ratio is defined as the outlet opening area divided by the inlet opening area. For example, a plenum A-Ratio of 2 w
38、ill decrease the low-frequency attenuation by 2 dE3, a plenum A-Ratio of 3 will decrease the low-frequency attenuation by 3 dB, etc. The linear reduction in attenuation is independent of the plenum size, wall lining type, and 1/3-octave band frequency. This relation- ship is only valid atff, Equatio
39、n 6 should be applied. Although the plenum orientation involving an A-Ratio 1 .O was briefly studied in this experi- mental work, it will not be reviewed in this paper. STATISTICAL LIMITATIONS OF THE ACOUSTIC MODEL The application of the proposed model may provide an accurate means to predict plenum
40、 TL when using similar wall absorption characteristics and panel materials found in this project. As shown in Figure 12, the new model (applying Equations 5 and 6) provides a significant low-frequency improvement over Equation 1. For plena applications within a practical sized envelope of 20-1 100 f
41、i3 (0.5-30 m3) volume and 50-650 fi2 (4.5-60 m2) surface area, using duct sizes in the range 12 f, This prediction tool was further improved through regression analysis, which consid- ered the terms of Equation 1 as a coeficient of attenuation (Kh. This analysis fitted the test data in the f 7 f, ra
42、nge to produce a good correlation among the extensive specimens used in this study. For the important low-frequency phenomena, Equation 1 does not apply. Direct correlations with the magnitude of plenum surface area (A-factor) and corresponding surface absorption characteristics (wall effect) were f
43、ound to be the key parameters in determining attenuation in the frequency bands f If,. A simplified approach, as shown in Equation 5, was found to accurately quanti the low-frequency TL for in- line plena. Significant TL increases, above those predicted by Equa- tion l, were found for varying offset
44、 angles for end-outlet plena. The elbow effect, created by a side-outlet configuration, produced a 2-6 dB increase in TL over a corresponding in-line configuration at all test frequencies except the 250 and 500 Hz octaves. In this range, a 0-3 dB reduction in TL was realized. This was typically the
45、region of TL peaks for the in-line configuration in this study. Therefore, the elbow effect may diminish the effective tuning of a simple expansion chamber or in-line plenum, as predicted by Equations 5 and 6. Due to the scatter of data, no definitive algorithm could be developed to accurately descr
46、ibe the elbow phenomena. The attenuation increase at the f f, range from an offset, end-inletlend-outlet plenum was found to be significantly higher than a corre- sponding elbow effect for an equivalent plenum geometsy. One or two outlets added to either an end-outlet or side-outlet plenum configura
47、tion produced negligible changes in broad- band attenuation, with respect to a referenced critical path used in the plenum system. Therefore, the user ofthe proposed model may ignore the addition of similarly sized outlets, within the size and quantity used in this study. As the outlet size increase
48、d, the low-frequency attenuation was found to decrease at a rate proportional to the ratio of outlet to inlet area. REFERENCES ASHRAE. 1999.1999 ASHRAE Handbook-HVAC Applica- tions. Atlanta: American Society of Heating, Refrigerat- ing and Air-conditioning Engineers, Inc. ASTM. 1999. ASTM C423, Stan
49、dard test method for sound absorption and sound absorption coeflcients by the reverberation room method. American Society for Test- ing Materials. ISO. 1996. IS0 9614-2, Acoustics-Determination of sound power levels of noise sources using sound intensiw Part 2: Measurement by scanning. International Organi- zation for Standardization. Munjal, M.L. 1987. Acoustics of Ducts and Muflers. New York: Wiley-Inter-Science, Inc. Wells, R.J. 1958. Acoustical plenum chambers. Noise Con- trol (July). 606 ASH
