Noise Reduction In High Efficiency Compressors

Transcription

1 Purdue University Purdue e-pubs nternational Compressor Engineering Conference School of Mechanical Engineering 2000 Noise Reduction n High Efficiency Compressors A. Faraon Electrolux Compressors P. Olalla Electrolux Compressors Companies Follow this and additional works at: Faraon, A. and Olalla, P., "Noise Reduction n High Efficiency Compressors" (2000). nternational Compressor Engineering Conference. Paper 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 Herrick/Events/orderlit.html

2 NOSE REDUCTON N HGH EFFCENCY COMPRESSORS Alessio Faraon Electrolux Compressors, R&D\taly, Via Consorziale 13, Pordenone, taly Phone: Fax: alessio.faraon@notes.electrolux.it Pedro Olalla Electrolux Compressors, R&D\Spain, Antoni Forrellad 2, E Sant Quirze Del Valles, Spain Phone: Fax: pedro.olalla-ayllon@notes.electrolux.es ABSTRACT This study deals with reducing noise in high efficiency compressors. Empirical and simulation techniques have been used, the former for analyzing and detecting the main noise sources and the latter for designing the modifications that have been applied. Two main compressor parts were analyzed and redesigned. The suction muffler was studied in order to reduce the high frequency noise caused by the movement of the suction valve, while maintaining good performance at low frequencies; and the shell was identified as a bad filter at high frequencies. The improvements described in this paper were applied to a full range of R134a and R600a compressor for household appliances. NTRODUCTON Today's market demands more efficient compressors every day, but with noise levels that are the same or lower than the noise levels of current compressors. When we try to improve the efficiency of a compressor, some of the modifications applied can increase the compressor noise level. This is the case when we remove the suction valve stop. When this is done, we improve the capacity and efficiency of the compressor, but we also increase its noise level and the excitation of other compressor parts, such as the shell. PRODUCT The objective was to increase the efficiency of a Rl34a and R600a range of compressors ranging from 60 to 210 Kcal!h, with displacements from 3.5cc to 9cc for R134a and form 4.5cc to 16cc for R600a, while keeping the noise level invariable for cooling capacities equivalent to the previous version. The efficiency increase was obtained by removing the suction valve stop, by modifying the valve plate and with a new electric motor. At the same time the suspension system was changed from the original suspended system to a bottom supported system. This modification implied using a modified shell while keeping the thickness invariable. Fifteenth nternational Compressor Engineering Conference at 957

3 NOSE LEVELS When all these modifications were applied, the efficiency objective was achieved, but the noise level was increased considerably to between 3dB(A) and 5dB(A), as shown in Figure 1, for a particular model and for a similar cooling capacity. The 1/3 octave band analysis shows that the frequency bands that most contributed to the deterioration ofthe overall noise level were the high frequency bands between 4000Hz and 8000Hz, but a sizeable increase was also detected in the range from 800Hz to 1600Hz n a preliminary analysis, two possible sound sources were thought to be responsible for this noise increase: );;> The suction valve movement, particularly for medium frequencies between 1OOOHz and 1600Hz and high frequencies between 4000Hz to 8000Hz );;> The shell, excited by forces transmitted through the suspension springs and the internal discharge tube in the frequency range from 1600Hz to 3150Hz. n order to reduce the noise level, it was necessary to improve or change the behavior of the principal acoustic filters of these two noise sources: the suction mufller and the shell itself HGH FREQUENCY NOSE Presuming that the movement of the suction valve was the main contributor at the frequency range of 4000Hz to 8000Hz, we decided to analyze the effect on the spectrum and overall noise level by reintroducing the suction valve stop. When the suction valve opening was limited by the stop the noise level was considerably reduced at these frequencies, and at the frequency band of 1250Hz (Figure-2). However, the compressor capacity was slightly reduced and, therefore, some efficiency was lost. This simple test led us to the conclusion that the suction valve was the main noise source at high frequencies and that the only way to minimize it, while maintaining the performances, was to redesign the suction mufller in order to reduce the high frequency noise. At the same time, it was important to maintain the previous mufller' s good response at low frequencies. New Muffler Design Taking advantage of the fact that the suspension system had been changed from a suspended one to bottom supported system and there were subsequently no limitations for increasing the muftler volume because the spring brackets had been removed, the muffler volume was increased by 77%, from 30 cm 3 to 53 cm 3, by extending the mufller on both sides (Figure-3). The design of the new muffier was studied to provide good performance at high frequencies, while maintaining a good response at low frequencies to avoid the excitation of cavity modes. The Fifteenth nternational Compressor Engineering Conference at 958

4 internal circuitry was modified from a single volume with two quarter wavelength resonators to a system with two expansion volumes and two Helmholtz resonators (Figure-4). The transmission-loss results for the two mufflers (Figure-S) show that the new muffler has a higher attenuation at medium and high frequencies compared to the previous one and also behaves better at low frequencies. This can also be concluded from the sound pressure measurements made at the suction muftler inlet (Figure-6). n these measurements, a lower radiation at high and medium frequencies can be observed when using the new suction muftler. With this new muftler the overall noise level was reduced by approximately 3-4dB(A), depending on compressor displacement. From the 1/3 octave band spectra we can see that the frequency bands affected by the new suction muftler are 1250Hz, 1600Hz and from 4000 to 8000Hz (Figure-7). SHELL MODFCATON When the new suction muftler was introduced the 2000Hz band became the dominant one in the spectrum and great dispersion at this frequency band was also observed. n order to further reduce the compressor noise level, the response at this frequency had to be improved Noise Radiation n order to better understand the compressor noise radiation at this frequency band, the compressor sound power level was measured by using the sound intensity technique. With this technique we were able to rank the noise emitted by different parts of the compressor. Only four areas were considered: the cover, the bottom shell and the front and back (Figure-8). After a preliminary measurement was taken, the right and left sides were not considered. From these spectra we saw that the bottom shell was the main radiator at 2000Hz. Despite the fact that the highest peak was at the 3150Hz band, due to noise radiated by the cover, it was not considered important because, in our sound power measurements in semi-anechoic room, it did not appear as a dominant frequency. Considering the bottom shell as the main radiator at 2000Hz, a more detailed analysis of this area was conducted. Two analyses carried out: Narrow band sound intensity map of the compressor bottom shell Experimental modal analysis of the compressor bottom shell From the results of these two analyses, it was seen that the main peak in the 2000Hz octave band happened at 2096Hz corresponding to the bottom shell's first natural frequency. t was also observed that the sound intensity map and the mode shape for this frequency had the same shape (Figures 9, 10 and 11). From these results it was concluded that the bottom shell's first natural frequency was the cause of the high noise at 2000Hz. Fifteenth nternational Compressor Engineering Conference at 959

5 Bottom Shell Modification We initially decided to design a new shell but, after bearing in mind that the bottom shell's first resonance was de-coupled from the rest of shell modes, the possibility of modifying only this part of the shell was considered and analyzed. The objective was to increase the stiffness of the bottom shell in order to move the resonances to higher frequencies and, particularly, to move the first natural frequency to a much higher frequency. n order to evaluate the impact of this increase in stiffness on the 2000Hz octave band, a shell was built with two welded ribs stiffeners on the bottom. This modification moved the first natural frequency from 2096Hz to 2448Hz, and produced a noise reduction of 15dB at 2000Hz and an overall noise level reduction of 3dB(A) (Figure-12). The next step was to redesign the bottom shell in order to get a response similar to that of the prototype developed with rib stiffeners. The new bottom shell was redesigned using FEM. A parametric geometric model of the shell was built and several solutions were simulated with ANSYS code. The solution with highest modes was considered the best and tested on the compressors. Table-1 shows the difference between the old and new shells for the same conditions (not operative). t can be observed that the frequency of the bottom shell's first two modes increased by more than 700Hz. OLD SHELL NEW SHELL Table-1 Table-2 shows the bottom shell's natural frequencies (full compressor) for the three shells used: the old one, the one with rib stiffeners and the redesigned shell. From these values we can see that the new shell meets the requirement of having a first natural frequency that is higher than the previous design. OLD SHELL SHELL WTH NEW SHELL RB STFFENERS 1 51 mode "dmode Table-2 With this new shell the noise reduction at the 2000Hz octave band was 8dB and the overall noise level was reduced by 2dB(A) (Figure-13). Fifteenth nternational Compressor Engineering Conference at 960

6 CONCLUSONS Two main noise sources, the suction valve and the bottom shell, were identified by means of experimental methods and these parts were redesigned and improved using simulation tools. The suction valve was characterized as a high frequency noise generator. Using an optimized suction muffler, this effect was minimized and the compressor performances remained invariable. The compressor shell was found to be the main radiator at 2000Hz. Using the sound intensity technique and modal analysis, the bottom part of the shell was identified as the main radiator at this particular band. nstead of redesigning the whole shell, only this part was modified by increasing its stiffuess in order to move its first natural mode to a higher frequency. With this modification, the noise level at this frequency was reduced by 8dB. ModifYing these two compressor parts (the mufller and bottom shell), the noise level of a entire compressor range for R134a and R600a was reduced by between 3dB(A) and 5dB(A). REFERENCES 1) Soedel, Werner, Sound and Vibration (Noise Control) of Compressors Short Course, Purdue University, July ) Hamilton, James F., Measurement and Control of Compressor Noise, Ray W. Herrick Laboratories, 1988 Fifteenth nternational Compressor Engineering Conference at 961

7 SOUND POWER.KVELS A WEGHTED,---, t t----t---+--'1--t-+--l- A \.-- r-r&r ' ( , / H n 16 :?'\» fv'l/ 1'-- 1- " V.. "',., SOUND POWER LEVElS A WEGirrF.D GQY75AA...) CA -[ -..A.! f' 10 \-- /--r-- \ 1\ # v--- f.../ DO lf:w MD nm :J,y)...W lOO llnlloo:tll:h:qa) v.t...\\'ith---+-satttdv.m. l Figure- 1 Figure- 2 Old Muffler New Muffler JUlL Old Muffler,] [' New Muffler L Figure- 3 Figure- 4 SOUNDPR levels MUFF..ERNLET :: -\ -1--1',1-\t-/\+-H f--t-t--H Jh'\11-h-!H ,.t t-+-+--f f------f-- --t-.:r.l\ <li.u.lullla-n<wmaiii<j-1 Figure- 5 Figure- 6 Fifteenth nternational Compressor Engineering Conference at 962

8 SOUND POWER LEVELS A WlllGH'!D GQYSAA ; ; -!j J _L_L L_L_L_L_L_L_L_L_L L_L L L-- _ --,-,-, ---, L_L_l_L_!_!_!_L_l_l_!_--l- 1 t ro--1. r-r-r-r-r-r-r-r-t_t_t_t_t_t_t_t_t_t_ ?-?-?-?-?-? Figure -7 S01JNJJ'O\V RUYH.S 1 -r-t-,-,-,-,--r-r-r-r-t-, 0 J -r-t-,-,-,-,--r-r-r-r-t-, -r-t--,-,--,--r-r _T_T l_l_ j -.---, rlfllllllllllltliiiu JDa..., i}di-sbdidj'-obd:! Figure ;--r-r- 80TTOM 5DELL SOUND rower : = ::: =!: :: =: L ini; = = : ::: : : ::: = = : ::: : : : 5 :3::E::::: :3::E:::::f:3::E:3:= "1- -t J-- -!U - -t-- '" t--; _:3:::::±:::::.. " :3::E===:==3::E:3::1::±: -, r t r ,- --- T - :J:: :::::: -J::c- :: ::r:::c:j::c:: T ,-- T -- ::::::=cj::c::: T-,--,-,--r-T- :::::::c:j::c:j: t t :: ::!:::c:j::c:!: 4Q C'7 Rl Figure- 9 Bottom Shell Sound intensity map (2096Hz) Figure- 10 # 1:2090 Hz z Bottom Shell Modal analysis (2096Hz) Figure- 11 Fifteenth nternational Compressor Engineering Conference at 963

9 n SOUND POWER l.f.vrlsa WF.G'F.D GQV75AA _L_L_L_L_L_L_L_L_L_l_L_l 1 SOJNDJ'OWERLEVKLSA WHGH'P.D GQY7SAA -r-r-r-r-r-r-r-r-r-t_t_t_t_t_t_t_t n _L_L_L_L_L_L_L_l_L_L_L_l L--l_l_l_- J w ' , ,-.-,-,-,-,-,-.-.-, l Figure- 12 ' 1 r-f-f-r-r-r-r---i-i_f_i_t_i_l_ w J- 1 s ,- -r-r-r-, r ! l l Figure- 13 Fifteenth nternational Compressor Engineering Conference at 964