Purdue University Purdue e-Pubs

International Compressor Engineering Conference School of Mechanical Engineering

1992 A Systems Approach to Appliance Compressor Quieting Using Active Control Techniques D. G. Smith Noise Cancellation Technologies. Inc.

M. F. Arnold Noise Cancellation Technologies. Inc.

E. W. Ziegler Noise Cancellation Technologies. Inc.

Kh. Eghtesadi Noise Cancellation Technologies. Inc.

M. Brown Americold

Follow this and additional works at: https://docs.lib.purdue.edu/icec

Smith, D. G.; Arnold, M. F.; Ziegler, E. W.; Eghtesadi, Kh.; and Brown, M., "A Systems Approach to Appliance Compressor Quieting Using Active Techniques" (1992). International Compressor Engineering Conference. Paper 821. https://docs.lib.purdue.edu/icec/821

This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact [email protected] 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 A Systems Approach to Appliance Compressor Quieting Using Active Noise Control Techniques

D.O. Smith, M. F. Amold, E. W. Ziegler, Jr. , Kh. Eghtesadi

Noise Cancellation Technologies, Inc. 1015 West Nursery Road Linthicum, MD 21090

M.Brown

Americold 2340 2nd Ave NW Cullman, AL 35055

Introduction

The purpose of this project wu to quiet a standard household mfrtgeratOr/frtezer using IICli.vc noise ICcluction rcclmiques. The model chosen, for demonsttation was US manufactured, had exposed coila and an open compressor compartment. The goal was to achieve the quietest l'tfrigeruor of its class that included a Japanese active cancellation model. Tho activity quickly focused on the single noisiest component- the compressor; however, the ovcnill goal of a feasible, quiet mfrigetator was the overriding conmamt.

Active noise attenuation, a concept mst described over fifty years ago, is artificially produced to counteract unwmted sound. The fundamental problem in active noise control in is to engineer the mixing of sound waves so that their superposition produces desuuctive intmfemncc throughout the whole sound field. This mixing of in-phase and uui-phue SOUDd a& the llUDe acouatie powc:r level as the original noise field auenuatcs the r.Uwu noise field.

Them IUC two main llppi'OICbes to ICtivc noise a&~enuadon. Tho first is to process the original sound and inject it back into the sound field in un:i-phasc. The second approach is to synthesize the canceling waveform with or without prior knowledge of the original SOUDd. Any active noise eonaol. application will consist of the following, :see Figure 1 primary noise source (co~~:~pn~ssor, fan, etc.), secondary noise source (speaker, shaker, en:.), detector(s) for primary noise and msidual error (accelerometet', , etc.) and electronic controller to minimize the electrical signal at the residual error sensor.

It has been demonstralCd that global attenuation of noise sources can be achieved only if the primary and secondary sources IUC located in close proximity (on the order of one-tenth of the wavelength of interest). 1f the secondary Source is several wavelengths distant from the primaly source, the room intcrfcrcncc pattern will be adjusted by changing the residual error microphone location. But, a minimum sound pressure level will sdll be achieved at the error microphone. However, the size of the quiet zone is only one-tenth of the wavelength of interest and not global.

317 In this paper, application of an active and passive noise control system for a household ICfrigmuor is evaluated. Experimental results showed that significant global ancnuation can be achieved using this control system.

Experimental Methods

The method of solutiOn to the problem was to rake data on the noise sources, investigate transmission paths and to design a solution based on the physical consiillints of the system. The solution could not increase the footprint of the appliance or cause operating problems with the compressor due to heat transfer.

Measurements were llbn on lhc compressor itaclf In a reverberant room. Both A-weighted and linear plots were~- The compressor was operated on a load stand at the nominal suction and discharge pressures of 19.2 and 175 psia at 115 V.

The initial data (plotted in the results section) showed both low frequency, below 1000Hz. and higher frequency noise from the compressor. The solution was therefOIC a combination of active and passive . Active was designed for lower frequencies and passive designed for the higher frequencies. Thus, an enclosure was built that Incorporated pusive and active noise cancellation.

In ordel' to control the refrigerator noise actively, Ncr used the synthesizing technique to genmrc the anti-noise wavefonn from the primary compressor -noise, see Figure 2. The synthesizing wavefonn was generaled using an acceletOtnCtcr that detected the fundamental motor tum rate on the compressor shell, about 58.7 Hz. A residual microphone was placed outside the enclosu:re and the speaker inside the encloswb in proximity to the cmrqnessor. The Ncr conaoller wu placed outside the · enclosure. Passive measures consisted of 1/l inch open cell foam inside the enclosure.-

The enclosure wu designed both from an acoustic and heat transfer point of view. The volume wu tuned to 60Hz to incJeue the acoustic coupling of the anti­ noise signal. Therefore, this incn:ued noise with the pulive enclosure was in tum was canceled with the active syswn. For heat trusfer, the enclosure had an inlet slit 1/2 by 9 inches ne~~r the CXJIIIPieSSor moundnc feet and an outlet-port in the back of the enclosure facing the wall ·

Next, measurements were made on two appliances Tiith factory installed compresSOl'S. One appliance wu to be the conaol unit and was not further modified. The other was the experiJI:Iental Wlit with the active ellClolure inttallecL The measurements were made in a seurl.-teverbcrut room. about 1 meier from front and rear, with a room volume similar to that expected in a lmlll kiu:hen.

In parallel with the above effunl, Ncr abo began Joaldnl into ways to quiet the compressor itself. Both passive and active meam weze lnva1iptM.

Experimental Results

The results an: seen in the followin1 Figmct.

Fi~ 3 shows the A-weighted spccuum up to 200Hz ofD COIIJPlasor in lhe reverberant room. Some peaks an: noted with a low froq1ICIIC1 bludJaDd hump. The telationships of the peaks to physical phenomena in the COIIIpiiiiCir wen DOt yet clear.

The next plot, Figure 4, shows compressor data linear -wei.pted 1IP to 2000 lU. Oearly, all the peaks an: harmonics of the turn rate of the~· about 58.7 Hz.

318- This result led to the approach of synchronizing to the motor tum rate and canceling that frequency to take the energy out of higher harmonics.

The next plots, Figures 5 and 6, were taken in the small semi-reverberant room from rear and front respectively. Figure Sa shows the narrowband spectrum of the unmodified refrigerator. Figure 5b is the modified with control off and 5c is modified with control on. Figure 5b shows how the higher frequencies are attenuated and the 58.7 Hz line is increased with the addition of the enclos~ while Figure Sc then shows how that low frequency tone is reduced with the active control. Figures 6 a, b and show similar c results for measurements made in front of the unit. Overa.ll.l/3 Octave measurements are seen in Figures 7 a and b. The overall reduction is 5 dBA in the front and 10 dBA in the rear with the measurements taken in the small semi-reverberant room.

Table's 1-3 are data taken in controlled environments at the manufacturer. Table 1 is the sound pressure data showing the increase in 60 Hz tone but overall decrease with the enclos~ while turning the active on reduced both 60 Hz and reduced the overall a total of 4 dBA. Table 2 shows similar sound power data with an overall reduction of 4.5 dB A. Finally, the environmental test at 110 degrees F was run to determine the effect of the enclosure on compressor temperature and efficiency. Table 3 shows the shell temperature stayed below the goal of 250 degrees F. Further heat aansfer design would lower this muimum shell temperature.

Conclusions

The problem of refrigerator quieting has been approached as a system. Overall system noise was reduced by a combination of active and passive techniques using elli.sting technology. The compressor itself was analy:zed and the application of quieting methods was researched with promising results. This feasibility demonstration has shown that a quieter system can be obtained without modifying the compressor itself. Future work in designing a quieter compressor combined with the active enclosure will produce a very quiet appliance.

References

1. Widrow, B., et al, "Adaptive Noise Canceling: Principles and Applications", Proceedings of the IEEE, 63(12):1692-1716, 1975.

2. Widrow, B. and S. Sterns, "Adaptive Signal Processing", Prentice-Hall, 1985: 3. Chaplin, G. B. B., "Anti-Noise -the Essex Breakthrough", Chartered Mechanical Engineer, pp. 41-47, 1983.

4. Glover, J., "Adaptive Noise Canceling of Sinusoidal Interferences", University, Stanford Stanford, California, Ph.D. Dissertation, 1975.

5. Nelson, P. A. and S.J. Elliott, "Active Control of Sound", Academic Press, 1992. 6. McLoughlin, M.P., Eghtesadi, Kh., Smith, D. G., and E. W. Ziegler, Jr., "Active Control of an Enclosun: with Multiple Transducers: Simulation and Experimental Results", Inter-noise 91, Dec 2-4, 1991, Australia.

7. Hamilton, J. F., "Measurement and Control of Compressor Noise", Purdue, 1988.

319 Figu1'e 1. Active Nonill!!l Controt Ba•ic Sy.t~:~M

Compre ..or

A.cceh,tomete1'

Figure 2. Active Refrigerator Enclasurl!l

s:=IU:=l GFIP SPEC UPPR VW eoiJB CH " FR 2{JOHZ DG - 10:18 WlG FT ' 2.V ~6 3B 09 SPEC A "'VG " BO ~-~ I j_ -··-- ·-·-- ·---· -·1 f---- I ;JBR I I

LOG .:11~ M 'VI'\ J1 . ... I··· ~-v\ ~·. ··- ·- - -

l~ ~ Vi)hA '\M ~ I ''WI .I I I'"' I' l 200.00 0 NORM LNX \ - BlSE oF .5000 H! SPEC SUM ~ iO 0 77 500 HZ •VG. 25. < OBR

Figure 3. A-Weigh~ Compressor Noise Dala

320 .Fl s:· G~l' SPEC JPPFI V~ SOOB CH A F'l1 2K'1Z :5 31 20 S~EC A A\'G aG i'ooe wrf •r •

:J Y"'0.C LNX 1 -1~ A.;~

Figure 4. Linear Weighted Compressor Noise Dara

--~~-~-~ '"

Ill~

10'

IO•

1Q·l

!Q·I ~~J~~~~~\~~lv~\

10'

Figure 5 a. Nanowband Roar Measurement of Unmodified Refrigerator

fqOS!:I ::: c--f-~------~-----

10-'~-

10'

ID'

lO IO 0 l 00 200 300 .q,o :500 600 700 800 900 1000

FigureS b. Narrowband Rear Measurement, Modified Refircerawr Control OFF

321 10)

IO'

JO·•

\O·":-~:::--:::::---:::::----::::----::::----:._-::._-,_,__~__J@ - ~ - ~ 0 ~ ~ ~ ~ ~ Modified Refrigerator Figure 5 c. Narrowband Rear Measurement, Control ON

\11: j ''''!. '"

]!'! l

Ill....,

\0'

1\l ~

IO'

- ~ ~ ~ - ~ ~ ~ @ of Unmodified RcfrigCl'lltor Figure 6 a. Narrowband Front M.,..su,.,mcnt

ffqOS9 10-1.

10·1

1o•

10·~

10"'

lO·'

10•

to•

600 700 800 900 1000 IO"" 0 100 200 300 .00 soo Modified Refriserator Figure 6 b. Narrowband Front Meu~~te~~~ent, Conti<>! OFF

322 \Ol r 10'1

\0'

Jo•

[Q"

10'

ta-lo a IOO 'ZOO 300 400 .500 1500 700 ROO 90

Figure 6 c. Narrowband Front Measurement, Modified Refrigerator Con1r0l ON

50------~ 45L. _____ .. ------···· -· ... ·------~---·---1 40------"" ·------· .

35,·--­ ------·-·· -- ·······------~---~---·-·-- <:( 30------­ .. -·-·· ---·. --- ·------.. - ---ll-I!E--11-.------i·----- r:Il 25J ----­ v 20~------·------t-----::,...---..-l~-.ll'l-I!IHIHIIHII-III-III-----1 15+-----~------~~---L~--;I~HIHI~Hin.--~ 10~1 -----m------~~--~~~~HE~~~~~ 5 r- -·----- 1- - --111--w---w-m IJlHil--m-m-~IHI!:--~ 0----W,--,---~~~~~~~.y~~Y-~--~1 dBA 25 40 63 100 160 250 400 630 1 000 20 32 50 80 125 200 315 500 800 1/3 Octave Bands 43 dBA 41 dBA 38 dBA 1m Original - Passive Only B Active and Passive

Figure 7 a. Refrigerawr Comparisons, Front

323 8

60~------~ ----·· ------·------·------. ·--·------·-----.------] 45'--- I 40~------··------···- --·------,_--- 1- __] 35------· -· ------______I ~ 3QL------::-·l-l-----ilh-IJ,_....--3--II----i ~ 2St----~------1---~---l-t~~~~._._ __~ "0 20------·!ll------11--1- ,- - 15+--- 1 ar--~~~-----11---.-•--IHJ}-fi...-lll-lfi-W-.m---llli-lll--lll------j 5 I § l 0+---~~~~~~~~~~~~~~~~--~1 dBA 25 40 63 1 00 160 250 400 630 1000 20 32 50 80 125 200 315 500 800 1/3 Octave Bands 46 dBA 46 dBA 36dBA Active and Passive B Original - Passive Only B

Figure 7 b. Refrigerator Comparisons, Rear

60 HzdBA Total dBA 39.3 Baseline 32.7 37.5 With Enclosure Only 33.2 35.6 Active on 29.9 Table 1. Sound Pressure Data

60HzdBA Total dBA 47.6 Baseline 35.4 44.9 With Enclosure Only 36.2 44.1 Active on 30.4 Table 2. Sound Power Data

Shell Tempezature Degrees F Baseline 221.8 249.7 With Enclosure Table 3. Shell Temperature Data

324