Purdue University Purdue e-pubs International Compressor Engineering Conference School of Mechanical Engineering 2000 Identification and Reduction of Noise in a Scroll Compressor J. K. Lee S. J. Lee D. S. Lee B. C. Lee U. S. Lee Follow this and additional works at: http://docs.lib.purdue.edu/icec Lee, J. K.; Lee, S. J.; Lee, D. S.; Lee, B. C.; and Lee, U. S., "Identification and Reduction of Noise in a Scroll Compressor" (2000). International Compressor Engineering Conference. Paper 1496. http://docs.lib.purdue.edu/icec/1496 This document has been made available through Purdue e-pubs, a service of the Purdue University Libraries. Please contact epubs@purdue.edu for additional information. Complete proceedings may be acquired in print and on CD-ROM directly from the Ray W. Herrick Laboratories at https://engineering.purdue.edu/ Herrick/Events/orderlit.html
IDENTIFICATION AND REDUCTION OF NOISE IN A SCROLL COMPRESSOR Jin-Kab Lee, Seung-Ju Lee, Dong-Su Lee, Byeong~Chul Lee, Digital Appliance Laboratory, Inc. Un-Seop, Lee Product Plant II, Inc. ABSTRACT Low noise of air-conditioners is a strong requirement from users. The main source of noise in an airconditioner is a compressor. Therefore, noise reduction in a compressor is quite significant as an element teclmology in air-conditioner field. Recently, scroll compressors are widely used, because they feature low noise, due to less pulsation of gas pressure, than that of rotary compressors. For reduction of noise, the source of noise must be identified. This paper presents analysis for the noise and vibration causes in a scroll compressor and shows a modification of top cap design for a low noise of a scroll compressor. INTRODUCTION As for air-conditioners for general domestic use, a great teclmological progress is being made in reducing energy consumption, improving pleasantness, and making environment-friendly. Recently, noise-reduction, one of the basic elements to realize pleasant air-conditioning, has turned out to be an important subject, where researches have been concentrated. To reduce tl1e noise of an air-conditioner, it is important to reduce the noise of t11e compressor. Scroll compressors continue to replace reciprocating teclmology in residential and small commercial air conditioning applications. Tilis paper is about a study to identify the noise source for effective noise reduction and presents results of a design evaluation of top cap directed at reducing noise on an HVAC scroll compressor. The structure of a 3Hp. vertical scroll compressor, the object for tllis exanlination of noise source, is shown in Fig. 1. Its diameter is 146mm, height about 400mm. It is driven with 60Hz, using R22 refrigerant. MEASUREMENT AND ANALYSIS Major noise source identification The noise spectrum of the scroll compressor used for tlris research is shown as Fig. 2. The test conditions of the scroll compressors are the ARl standards, and tl1e noise has been measured witl1 the sound power of ISO 3744 standard. In Fig. 2 is shown a wide band of noise spectrum from O.SkHz to 4.5kHz, and the main peaks exist in tl1e bands of0.8khz, 1.6k-2kHz and 2.5kHz. At 1 meter from the top cap, tl1e noise spectrum oftlris scroll compressor 1041
is shown in Fig. 3. The total noise level is lead mostly by the peak at the 2. 7kHz range. Tins research makes a priority investigation of the noise source of the main peak frequencies of 2. 7kHz range shown in Fig. 2 and Fig. 3. We measured sound intensity to investigate noise contribution of the components assembled inside from tl1e direction of tlle noise' location and frequency radiating from the scroll compressor. 145 cylindrical virtual lattice were made, and the sound intensity was measured with a sound intensity probe (B&K3520) installed 10cm away from the compressor in 3 axes' directions, which was analyzed through LMS CAD A-X. Fig 4 shows contribution of l/3 octave band frequencies among the noises radiating from each part of the compressor. We can see tllat noise of 0.8kHz band occurs in tl1e suction part and the motor, l.6khz in the compressing part, 2kHz in the motor, and 2.5kHz-3.15kHz in the top cap. Fig. 5 shows intensity map of 2.5kHz band. We find that tlle largest radiation of 2.5kHz band occurs in tl1e top cap. To understand characteristics of the top cap of the compressor, which cause noise, we performed a experimental modal analysis and analyzed by means of ANSYS. A brief comparison of the results is shown in table 1. TI1e first resonance of the top cap appears in 2722Hz. With reasonable verification of the analytical model, it can be used to predict tl1e effect of design changes on modal frequencies and shapes. In order to obtain exciting force information due to pressure fluctuation, the frequency composition of the pressure was measured by installing a pressure sensor inside tl1e top cap. Fig. 6 shows the result. The exciting force by gas pressure is concentrated mostly under 700Hz, but a peale tl1at seems to be related to the resonance of the compressing chamber can be seen at the 2.7-2.8kHz range. The first resonance frequency of the top cap is around 2.7kHz, which coincides with the pressure pulsation frequency of 2. 7kHz-2.8kHz, causing high-level noise. So, tllough it is ideal to design tl1e first mode frequency oftl1e top cap at over 4000Hz, you can see that it must be designed at least at over 3500Hz. Design parameter study The factors tlmt affect tlle resonance frequency can be considered as the tlriclmess of the shell, the diameter of the curvature from tl1e center of tl1e top to the exhaust part, tl1e height of the top cap, the opening amount in the exhaust port area. In order to evaluate the effects of thickness, changing the conventional 3.2t to 4.5t, and also the 4 ribs installed on top cap (Fig. 1) was removed performed tl1e analysis. Fig. 7 shows the analysis result, and tlle ribs decrease tlle rigidity thus decreasing tlle first mode by about 350Hz. But as it proceeds to lrigher modes, it gave almost sinrilar results as the conventional form. The increase in frequency related to the thickness is about 350Hz. A uniform increase in frequency occurred in all modes. According to the above analysis, in order to raise the first mode frequency to over 3500Hz, tl1e tlrickness and form must be changed simultaneously. The result of calculating the resonance frequency by changing tlle curvature diameter from t11e center of the top to tlle exhaust port is shown in Fig. 8. TI1e curvature diameter could raise t11e resonance frequency by 800Hz. It also showed that over a certain value, tl1e increase of frequency was mitrlmal although tl1e diameter decreased, and the frequency alteration decreased as t11e modes increased. The height of the top cap is related to the top curvature diameter, so tlle change of resonance frequency was obseived by increasing the cylinder part of the edge. Results of 1042
the analysis (Fig. 9) show that as the height increases, the resonance frequency decreased. Since the increase of length is related to the total length of the compressor, the lower the better. But the change of pressure inside the compressing chamber must be considered also. As you can see above, by calculating the resonance frequency in many angles, the most sensitive factor is the curvature diameter R of the top cap. Other factors such as thickness and height affect similarly, and the angle of the opening in the exhaust port doesn't affect it greatly. So, in order to design a top cap with a high resonance frequency, the curvature diameter of the top must be set small and the height short. The opened angle doesn't greatly affect it, but it is advisable to set the angle at U1e maximum resonance frequency after the curvature diameter is set. Design evaluation Based on the above results, we tried to check tl1e affects of increasing resonance frequency by designing the top cap so that the first mode exceeded 3500Hz. The form of the newly designed top cap is shown in Fig. 10. Fig.ll illustrate the analytically determined mode shapes for the first and second modes. The results of impact response type experimental modal analysis between tl1e original and new top cap are shown in Fig. 12. The first resonance frequency measured to be 3.8kHz. As you can see from the above results, the resonance frequency of the conventional top cap at 2.7kHz matched with the resonance element, Ums the noise level at this frequency range was very high. But the noise of t11e improved top cap with the same tllickness bas decreased by about 15dB in the 3kHz range by changing the form to raise the rigidity. The sound power level of the whole compressor has decreased at an average of 3dB. In Fig. 13 sound pressure levels are compared for the original top cap and the new top cap. CONCLUSION In this paper, experiment analysis was performed to identify t11e main noise sources and the results of redesigned top cap with the goal of lowering tl1e noise radiation were shown. The results have been concluded as follows. 1) It was determined that a large contributor to overall compressor noise was the top cap/discharge plenum assembly. The resonance frequency of 2. 7kHz range at tl1e top cap gets excited by the pulsation of discharge gas, causing high-level noises. 2) The factor that mostly affects the resonance frequency of the upper shell is the top curvature diameter, then the tllickness and then the height. Others hardly affect it. 3) The increase of resonance frequency according to the tllickness shows a mliform increase in all modes. But the affect according to the curvature diameter is great in earlier modes, but not much as t11e modes increase. 4) Since theoretical analysis and test results match, by using this research method, a top cap with low noise radiation can be developed. Fifteenfu International Compressor Engineering Conference at 1043
REFERENCES (I) Y.H.Cho, B.C.Lee, J.K.Lee, "Development of High Efficiency Scroll Compressors for Package Air Conditioners", Proc. Purdue Compressor Conference, 1996, pp.323~328. (2) M. Shuji, N. Shiniji, "A Study on noise reduction in a scroll compressor", Proc. Purdue Compressor Conference, 1996, pp.605~10. (3) Sano, K., Kawahara, S., Akazawa, T., Ishii, N., "Experiment Study for Reduction of Noise and Vibration in Hennetic Compressor (1st Report: Reduction of Noise caused by Resonance of each Scroll Compressor Element)", Trans. of the JSRAE, 1997, Vol. 14, No.2, pp. 125-136. Fig. 1 Cross-sectional view and original top cap of scroll compressor db{a) 1000 2000 3000 4000 5000 6000 Frequency(Hz) 1000 2000 3000 4000 5000 6000 Frequency{Hz) Fig. 2 Typical noise spectrum Fig.3 Noise spectrum at top cap 1044
70 r------------------------------------. 60. (.: ::._.=.-: ~ -:- ;;:~:. -.. _... _.. ::;.t. ~-..,.., so c :. :~~:-"""' 40 _.. -Top cap Suction Motor - - Sub frame 30 ll) 0 ll) C\1 0 C\i 1/3 octave band(khz) Fig. 4 Sound Intensity pattern at each part ll) C\1 Fig. 5 Sound intensity map at 2.5kHz band Fig. 6 Pressure pulsation in discharge cavity Table 1 Modal frequencies of original top cap Mode Calculation(Hz) 1 2946 2 3944 Experiment(Hz) 2722 3548 Error(%) 7.6 10 1045
!- -CASE I -EJ- CASE II --.--case 1111 8000 7000 i 6000 ~ :ii 5000 :::> J: 4000 0" 3000 2000 -- --------------~----------------- ---------------------- p 2 3 4 Mode ------------~----------------------- -'------------------------ 5 6 Fig. 7 Changes of modal frequencies according to thiclmess and Rib CASE I: Original top cap (3.2t), CASE II: Original top cap without rib (3.2t), CASE III : Original top cap ( 4. 5t) [ 0 R=ao -&-R=330 -+-R=350 ~R=400 I 'N I 8000.--------------------. 7000 :;: 6000 u c 6-5000 Q) Q) Li:: 4000 --------EF' --------~-0'... --- 3000 L..---------'------------ 2 3 4 5 6 Mode Fig. 8 Changes of modal frequencies according to curvature diameter 1-- 3 H==72.1 ~H=49.71 9000..--------------------..., 8000 i 7000 ~ t 6000 :::> J: 5000 rr - - - o 'EJ..... 4000 3000........... 2 3 4 5 6 Mode Fig. 9 Changes of modal frequencies according to height 1046
Fig. 10 New top cap :first(3803hz) second(3959hz) Fig. 11 Mode shape at each frequencies db(a) Frequency(Hz) --- NewT/Cap... Original T/Cap Fig. 12 Comparison offrf between original and new top cap 70 60 ~ 50 c::j Original co -a 40 llllll!illlllll New 30 20.r l.r.i. lll.fl. IJ '.rln 200 400 800 1600 31 50 6300 12500 A Frequency(Hz) Fig. 13 Comparison of 1/3 octave band sound pressure spectrum between original and new top cap 1047
1048