DECEMBER 15-18, 1997 ADELAIDE, SOUTH AUSTRALIA. Department of Mechanical Engineering, Hanyang University. Haeng Dang-Dong, Seong Dong-Ku, Seoul, KOREA

Transcription

1 FIFTH INTERNATIONAL w CONGRESS ON SOUND DECEMBER 15-18, 1997 ADELAIDE, SOUTH AUSTRALIA AND VIBRATION A design echnique for reducing he inake noise of a vehicle Jae-Eung Oh, Kwang-Hee Han Deparmen of Mechanical Engineering, Hanyang Universiy Haeng Dang-Dong, Seong Dong-Ku, Seoul, KOREA ABSTRACT : The inake noise of a vehicle is analyzed and reduced by using he ransfer marix mehod as well as he FE analysis code, NIT/SYSNOISE. A speaker exciaion sysem is also proposed, which can analyze he acousic characerisics of he inake sysem. I is easy o analyze he inake noise in he laborao~ environmen, and can be used a an early design sage of he inake sysem developmen. And his sudy proposes he improvemen o reduce he level of he inake noise. I is o selec he opimum posiion of a resonaor and verified by NIT/SYSNOISE, FE analysis commercial sofware and esing he prooype and a proposed speaker exciaion sysem. 1. Inroducion The inake noise, a major source of he vehicle noise, was sudied o reduce he level of he noise for many years, since he quie passenger comparmen was noifkd as an imporan qualiy for a vehicle. The inake noise is a low fkquency noise up o abou 600 Hz and a facor no only iecing he exerior bu also causing a booming noise in he inerior. Tradiionally, he inake noise has been analyzed and improved by he road es or he ~~ x fi&~@& ~,mr~fi fil ~ f&l ~r~d~~,. h& 4L!fdI d ~1 ~1 a~~%~%i?l. These approaches require very high cos, long ime consuming and he large size laboraory environmen and equipmen. Recenly, he various soflware ools and he analysis echnique have been proposed o analyze and improve he inake noise. Specially, he ransfer marix mehod and he acousic finie elemen analysis have been considered as usefil hings a he early sage of developmen of inake sysems. The finie elemen analysis is more useilbl a he concep sage of he developmen of he inake sysem. Bu once presened wih a deail design problem FE analysis is no a pracical ool anymore. This is because he relaive high cos and analysis ime are required in re-modeling and gaining sufficien accuracy, while he ransfer marix mehod is easy o analyze he inake sysem bu his mehod is applicable for he only simplified model. So he eehnique is usefil a concep design sage of a inake sysem.

2 And he simulaor mus be used o veri& he acousic characerisic of he proposed model by various analysis echniques. In general, he inake noise has been analyzed by he real road es or he experimen using he engine dynamo. Anoher mehod is roaing he crankshaf wih an elecric moor, bu i has some dkadvanage ha he analysis requires he large power moor and generaes addiional moor noise conaminaing he objecive one for a inake sysem. And he oher one ha pu a speaker in each cyliider insead a pison and excies he air inake sysem wih same phase dflerence as he firing of a real, was proposed. In a real roaing engine, when a cylinder valve is opened, res valve are closed, herefore he miolds of closed valve is operaed as he quarer wave resonaor. So his mehod is no expeced he exac analysis. In his sudy, he esing apparaus of he inake sysem which can use in a laborao~ environmen wihou addiional noise and he exhaus gas, is proposed. Tha has a speaker in only one cylinder of opened valve insead a pison and res valves are acually closed by he uned camshaf angle. The characerisic of he inake noise is like as a real. This sudy analyzed he real inake sysem using he ransfer marix mehod and FE analysis, NIT/SYSNOISE and he proposed model o selec he posiion of a resonaor. And hese are verified by he speaker exciaion sysem. 2. Speaker exciaion sysem for he analysis of he inake noise The speaker exciing sysem is proposed using a curren four-cylinder 1.8 liers engine. The crank shaf, connecing rods and pisons are removed from he engine and hen a speaker is se ino he lower dead cener of a cylinder. The cylinder wih a speaker is seleced by he esimaing his sudy selec he cylinder mos fhr from inle hose. The camshaf has four cams in his case. Cams are ied on he shaf wih uned angles. The inake valves are driven by he cams on he shaf. While a inake valve is opened, res valves are closed. The closed manifolds of inake sysem operae as he quarer wave resonaor (abou 170Hz and 180 Hz in his case). Analysis frequency range are 1000 rpm-5000 rpm ( abou 35 Hz -165 Hz ) and he exciing fi-equency incremen is 150 rpm (5 Hz). The order racking of he frequency response and he overall level are measured by he sound level meer a each ccdion. The exciaion is derived by he signal generaor and he amplifier, and he exciaion frequency is conrolled by Eq. (1 ) fieq.(hz)= nxpnxcycx-& (1) where, Jieq. is he exciing frequency, n is he harmonic order, rpm is he revoluion of he engine and cyc is he number of cylinders. Fig. 1 shows he experimenal configuraion of he inake sysem for he vehicle using a speaker exciing sysem. The speaker has 3inches dkmwers and 3 was power, which is excied by a signal generaor and an amplifier. The sound level meer is used o measure A- weighing noise sigmds. The signals are analyzed by a specrum analyzer and a compuer. The measuring poin of he inake noise is locaed 45 degree and 0.1 m verically from he inake enrance. In Fig.2, he normalizd sound pressure level obained from he real engine is compared wih ha from he speaker exciaion sysem for he verificaion. Two ses of he daa are relaively in good accord.

3 hldmmlnibld -~ I* how sound Immlmou / S&nkaku,7 (Inld, MA320)!2!? ml an FFr Mayzar (SA390) El Fig. 1 Experimenal configuraion using acousic exciaion sysem wih speaker u o.% 0.s o.a s 0.7 O.a :re.dc4fgim s _ :Sim&or 0.s1 L o mm 4000 sow Fig.2 Normalized SPL of he simulaor and he real engine a each exciaion rpm Fig. 3 % W 5000 fpm Sound pressure overall levels of a simulaor a each rpm The es inake sysem has wo resonaors: he uning hquency of main resonaor is 87Hz and he frequency of he auxiliary one is 75 Hz. Fig. 3 shows he acousic characerisics of he hree cases: 1) he case wihou resonaor, 2) he case wih a main resonaor and 3) he case wih boh resonaors. I is shown ha in he case wihou resonaor he sound pressure level is relaively dominan a abou 2700 rpm (87 Hz). This means ha he resonance is caused by

4 he excied air colwnn in inake sysem a 87Hz. The main resonaor is insalled o reduce he noise of ha, bu he insallaion of he main resonaor cause he resonance a 2000 rpm-2500rpm. And he auxiliary resonaor is insalled o improve he resonance of 75 ~ bu cause he resonance a 1700 rpm-2100rpm. Fig. 4 shows he order racking of he inake noise using he speaker exciaion sysem and he resonance of 70 Hz uncorrelaed wih he exciaion frequency. Since his resonance is in he general usage revoluion of he engine, he improvemen mus be needed. 1 $,3 1 Ooo Fig. 4 The order racking of inake noises a each exciaion rpm 3. Analysis of inake noise using he convenional mehod The ransfer marix mehod and he finie elemen analysis using NIT/SYSNOISE are used o ver@ heoreically he previous analysis resul and o propose improvemen for selecing he opimum posiion of he main resonaor. Fig. 5 shows ha he ransmission losses are obained by he ransfer marix mehod and he experimen using he wo microphones mehod. I is shown ha he predkion using ransfer marix mehod is well agree wih he measuremens; herefore he ransfer marix mehod can be used in he analysis, he design and he improvemen for he inake sysem..10; I Im Soo Fmq.(Hz) Fig. 5 Transmission losses of he inake sysem The FE model and he analysis resuls of he inake sysem is shown in Fig. 6, where he ransmission losses by NIT/SYSNOISE are nearly he same as hose of he ransfer marix

5 mehod. In his sudy, as he main resonaor is moved from he inake enrance o he air cleaner wih 5 cm incremen and he ransmission losses are compared a each posiions using he ransfer marix mehod. Fig. 7 shows he analysis resuls. I is shown ha he more close he resonaor move o he air cleaner, he less he resonance of 75Hz is, bu ha make he resonance a 110Hz. Since he noiseof11ohz(33oo rpm) is generally in high speed range, he noise is heard relaively smaller han low frequency noise, and also he booming noise is no caused. m 70 TL(U6A) Fig. 6 Esima~ 0 60 lmlm.2m2602m FREQUENCY (iz) TLs of he inake sysem using NIT/SYSNOISE The locaion of a resonaor is generally deermined on he basis of he pressure ani-modal poins of he mode shapes. Maximum aenuaion is achieved a or near he ani-modal poins. The acousic mode shape analysis using NIT/SYSNOISE is peflormed o find ani-modal poins of 87 Hz. Fig. 8 shows he FE model and he acousic mode shape of 87 Hz. I is shown ha he ani-modal poin is a he air cleaner and his analysis resul is he same as ha of he ransfer marix mehod. Based on hese analysis resuls, his sudy proposes he modified inake sysem wih a new locaion of he main resonaor on he air cleaner. 50n : :1 40 :1 :4 2= L -10 o 100 m m Fmq.(liz) Fmq.(iz) (a) #1 -#4 (b)#5 -#7 Fig. 7 Esimaed TLs a each posiion of a main resonaor using he ransfer marix mehod

6 -.-.. r=m.u.- l ,- IL (a) FE model Fig. 8 - L (6) acousic mode shape (87Hz) FE analysis using i;sysnoise Fig. 9 shows he FE model and he ransmission loss obained by NIT/SYSNOXSE. Fig. 10 shows he overview of he prooype of he proposed inake sysem. Fig. 11 shows comparison of he ransmission losses obained by he ransfer marix mehod and he measuremen using he prooype. Similarly wih he previous predicio~ he resonance of 75Hz is improved considerably, and resonance of 110 Hz is caused. L J...L...i...i...l...i...l... U,.-= mmmae Fig.9 (a) FE model ( b)rm=slon loss FE analysis of he modified inake sysem using NIT/SYSNOISE The sound pressure levels are measured o verifi he acousic characerisic of he proposed prooype using he speaker exciaion sysem. Fig. 12 shows ha he inake noise characerisic of he modified sysem wih he new locaion of a resonaor are compared wih hose of he original sysem wih a main resomor as well as he original sysem wih all resonaors. The order racking of he modified inake sysem is shown in Fig. 13.

7 Fig. 10 Overview of a prooype of an inake sysem -1o I o 100 m m 400 Soo Frea.O+Z) Fig. 11 Transmission loss of an inake sys& wih modified posiion of a resonaor #..-.,..../4..!llc-EEz2 looo15002ax) z5w cm Fig.12 Sound pressure overall level a each exciaion rpm The inake noise of he modied sysem wih only one resonaor is lower han ha of he original wih wo resonaors on he whole range excep a 3300 rpm(110hz). And i is known ha he more he revoluion speed of he engine increase, he louder he inake noise is heard. If his endency is considered, he inake noise of he modified sysem is superior o ha of he original in he side of he sound qualiy.

8 ,: Fig. 12 The order racking of he modified inake sysem 40Conclusions The speaker exciaion sysem is proposed, which can analyze he acousic characerisics of he inake sysem and be used a he early design sage of he developmen. The inake noise of a curren inake sysem is analyzed by he ransfer marix mehod, he experimen using he wo microphones mehod and he finie elemen analysis using NIT/SYSNOISE. Improvemen o selec he opimum locaion of a resonaor is also proposed o reduce he inake noise efficienly by he compuer simulaion. The improvemen is verified by he prooype and he speaker exciaion sysem. Refermwes (1) Takeshi TOI and Nobuyuki OKUBO, Noise Predicion of Air Inake Sysem for Vehicle and is Applicaion, Iner-noise 95,1995, pp (2) Yoshialca Nishiro and Tokio Koh~ Osamu Kurod~ New Approach o Low-noise Air Inake Sysem Developmen, SAE , 1991, pp (3) SYSNOISE manual, NIT (4) M.L.Munjal, Acousics of ducs and mufllers, John Willey & Sons, 1987 (5) John D.Kosun and J.S.LQ Effec of resonaor locaion on resonaor effecivenessusing NASTRAN mode shape predkion analysis and LAMPS acousic model, SAE , 1994, pp. 1-12