Journal of the Korean Physical Society, Vol. 55, No. 6, December 2009, pp. 2405 2410 Spectrum of Infrasound and Low-frequency Noise in Passenger Cars Sung Soo Jung, Yong Tae Kim and Yong Bong Lee Korea Research Institute of Standards and Science, Daejeon 305-340 Ho Cheol Kim Korea Advanced Institute of Science and Technology, Daejeon 305-701 Su Hyun Shin and Cheolung Cheong Busan National University, Busan 609-735 (Received 14 July 2009) The interior sound pressure spectrum of modern passenger cars running at various speeds on various road conditions was analyzed in terms of a 1/3 octave band over a frequency range of 1 to 250 Hz. The car speed varied from standing idle to a motorway speed of 110 km/h on a flat asphalt road and a cement road and in a tunnel. The 1/3 octave spectrum consisted of a nearly flat sound pressure level in the infrasound region up to 10 Hz, followed by a gradual decrease with frequency. The spectrum contained several humps, which corresponded to the resonance frequencies of engine firing and rambling noise. The spectrum profile suggests that much of the noise energy is in the infrasound region and increases with car speed. The levels of infrasound and audible sound pressures in modern Hyundai cars were shown to be significantly lower than those in the equivalent cars of 1970 models due to the incorporation of new materials with improved structural design. PACS numbers: 43.28.Dm, 43.50.Lj Keywords: Low-frequency noise, Infrasound, Sound pressure level, Passenger car DOI: 10.3938/jkps.55.2405 I. INTRODUCTION Recent development in car design incorporating with the use of advanced materials for structural components are making significant improvements in all aspects of performance, including noise quality. Aspinall [1] initiated the works on periodic noise of wind throb and rattling by traveling cars to understand the mechanism of such noise generation with a view of helping vehicle designers reduce the effects. He found that transportation noise includes airborne combustion sound and mechanical/aerodynamic noises of inaudible infrasound of 20 Hz and audible vibration and sound of between 20 Hz and 20 khz. The sound pressure level inside the car exceeded 120 db when the speed of the vehicle was varied. This is coupled to the interior of the car, which with an open window acts as the neck of a Helmholtz resonator [2]. Tempest and Bryan [3] extended the frequency range of the sound pressure level to the octave band over a frequency range of 2 Hz to 16 Hz to establish the levels of infrasound to which passengers and drivers are subjected in the passenger seat. The microphone recorded for a period of approximately 1 minute for a speed range from 64 E-mail: jss@kriss.re.kr -2405- km/h upwards. Generally, the pattern of the spectrum reported indicates that a nearly flat spectrum for 2 32 Hz is followed by a progressive reduction in octave band levels at high frequencies [3, 4]. The infrasound is also thought to be a hazardous environmental issue. Extended exposure is known to have a physiological impact on the human body and psychological trauma via neurological interference, including medical disorders, such as nausea, haziness, disorientation, and potential internal organ damage as the amplitude is increased [5-9]. The 7-Hz infrasound is believed to have the most profound effects and is commonly associated with the resonant frequency of the body s organs. Since the early works in the 1970 s, further work available in the open literature is rare for some reason. However, the noise pollution arising form traffic systems has become a serious social problem with escalating increase in the number of cars and speed on the road. Recently, Onusic and Mizutani [10] reported an experimental work on the infrasound pressure levels in commercial vehicles and passenger cars. Microphones were positioned near the driver s ear, and acoustic data were obtained using a FFT with a constant and narrow bandwiths of mhz above 4 Hz. The sound pressure level (SPL) in the medium passenger cars of the Santana driven at 120
-2406- Journal of the Korean Physical Society, Vol. 55, No. 6, December 2009 Table 1. Classification of passenger cars. Car Model Engine capacity Fuel type Cylinder Kilometer (cc) (km) number run (km) Avante 1495 gasoline 4 79000 NF Sonata 1998 gasoline 4 20000 Santa Fe 1991 diesel 4 78600 Equus 3342 gasoline 6 53000 km/h showed a number of spectral humps of 85 db for frequencies 20 Hz, then decreased with increasing frequency with a sharp hump of 78 db at 38 Hz. The SPL increased more than 20 db for an open left window compared to a closed window while driving at 120 km/h. The present work is concerned with the interior noise spectra of infrasound and low-frequency noise in modern small passenger cars at various car speeds ranging from idling to motorway speeds under different road conditions. The noise spectra of modern cars are also compared with those published in the last decade to see the improvements made in noise reduction. II. EXPERIMENT The tested passenger cars of Korean Hyundai Co. are listed in Table 1 with their characteristics. They have automatic transmissions and spring coils. The road tests were carried out by varying driving speeds from 80 to 110 km/h and idling on flat asphalt and cement roads, and for ascending and descending slopes. The interior noise of cars was measured with a precision sound level meter (B&K 2231) with a low-frequency condenser microphone (B&K 4193) and was recorded on a digital recorder (Sony PC 208Ax) at the middle of the rear seats. The noise spectrum was analyzed with a spectrum analyzer (B&K 3550) for the narrow band and a real-time frequency analyzer (B&K 2144) for the 1/3-octave band spectrum over a frequency range from 1 to 250 Hz. III. RESULTS AND DISCUSSION The noise transferred to the car body and interior has the characteristics of engine type and mechanical vibration of rotating tire/road and of aerodynamic sources varying with capacity, car speed, and age of the car. Tire conditions, such as groove geometry and depth, and tire pressure are also important variables affecting the noise characteristics. 1. Noise Spectrum of Idling and Driven Cars Fig. 1. Noise spectra of an Avante driving at 100 km/h on a flat asphalt road with a constant engine speed of 2500 rpm: (a) narrowband SPL spectrum and (b) 1/3-octave band SPL spectrum. The narrowband SPL spectrum of the Avante at idle, engine speed 2500 rpm in neutral gear, is shown in Fig. 1(a). The SPL was fluctuating 43 db on average at frequencies < 15 Hz and decreased to 23 db at 100 Hz as the frequency increased with pronounced spectral peaks of 75 db at 22 Hz and 65 db at 43 Hz and 87 Hz. All these spectral humps are identified as resonant frequencies of the engine rpm because the first resonant frequency of the 4-cylinder engine for an engine speed of 2500 rpm is about 84 Hz and the first and the second sub harmonics of 42 and 21 Hz are well matched to the measured frequencies of 43 and 22 Hz in the narrow band spectrum at the idle condition. The 1/3 octave band SPL spectrum of the same Avante driving at 100 km/h on a smooth asphalt road in Fig. 1(b) is markedly different form that at idle with the same engine speed of 2500 rpm. A nearly flat SPL of 90 db for frequencies < 4 Hz of the infrasound region was followed by a gradual and smooth decrease to 65 db at 250 Hz. The spectral humps at 40 Hz and 80 Hz corresponding to the center frequencies of the 1/3-octave band spectrum of the running car in Fig. 1(b) are identified as the sub-harmonic and the first harmonic, respectively. The spectral humps
Spectrum of Infrasound and Low-frequency Noise in Passenger Cars Sung Soo Jung et al. -2407- Fig. 2. 1/3 octave band spectra of an Avante tested on asphalt and cement roads: (a) running at 80 km/h and (b) at 110 km/h. at 22 Hz in the narrow band of the car at idle is close to the second sub-harmonic, but was not found in the 1/3-octave band spectrum of the driven car. The hump at 12.5 Hz in the 1/3-octave band is known to be due to the rambling noise or engine rotation unbalance. 2. 1/3 Octave Band Spectrum of Cars driving on Open Flat Asphalt and Cement Roads The 1/3 octave band sound pressure spectra inside an Avante driven on flat asphalt road at 80 km/h and 110 km/h are compared in Figs. 2(a) and (b), respectively. The physico-mechanical properties and asperity geometry in terms of height and spacing are known to be markedly different between asphalt and cement roads. A greater friction noise was induced on the tire/cement than on the tire/asphalt and is considered to be the main contribution to the higher SPL. The spectral humps at 40 Hz and 80 Hz correspond to the resonant harmonics, similar to those in Fig. 1(b), for the Avante running at 100 km/h. Fig. 3. 1/3 octave band interior noise spectra of cars running on an asphalt road: (a) Equus and (b) Santa Fe. 3. Noise Spectra of Modern and 1970 s Cars The 1/3 octave band sound pressure level spectra of the Equus (3342 cc, gasoline) and the Santa Fe (1991 cc, diesel) running at 80, 90, 100, and 110 km/h on a flat road are shown in Figs. 3(a) and (b). The sound pressure level of the Santa Fe (sports utility 4-wheel-drive vehicle) is higher than that of the Equus. The sound pressure level of the Santa Fe running at 110 km/h decreased from 96 to 89 db between 2 and 10 Hz, followed by a second hump at 80 Hz corresponding to the rpm resonance. The decrease in the sound pressure level with frequency was more gradual with a much reduced hump as the car s speed decreased. The interior sound pressure levels for frequencies < 95 db in the infrasound region we obtained from the Equus and the Santa Fe running at 100 km/h on a flat road in Fig. 3 are much less than the reported sound pressure levels of a 3-liter car and 2-liter wheeldrive utility vehicle of Tempest and Bryan in 1972 [3]. 4. Variation of Interior Sound Pressure Level with Car Speed
-2408- Journal of the Korean Physical Society, Vol. 55, No. 6, December 2009 Fig. 4. Changes in total sound pressure levels of infrasound of an Equus with increasing car running speed. Fig. 6. 1/3 octave band spectra of an Equus running at 100 km/h on flat, ascending, and descending roads. Fig. 5. 1/3 octave band spectra of an Equus running at 100 km/h on a flat asphalt open road and in a tunnel. Fig. 7. Sound pressure level of an Avante with different fuel types. The overall interior sound pressure levels in the infrasound region of the Equus running at speeds between 80 and 110 km/h are presented in Fig. 4. The noise levels in the infrasound frequency region of the Equus increased as the car speed increased. The sound energy (E) increases with the car speed (V) with index α, such that E(p 2 ) V α. The index α can be determined from Fig. 4, where α is found to be 5.8 for the Equus. 5. Noise Spectra for a Flat Road and a Tunnel The interior sound pressure level of the running Equus decreased in the tunnel from 99 db to 83 db at frequencies between 1 Hz and 6.3 Hz while that on the open flat asphalt road remained nearly steady at about 87 db on average, as in Fig. 5. The rapid increasing sound pressure level in the < 3 Hz region in the tunnel may be due to a tunnel resonance. The spectral humps of 87 db at 12.5 Hz both in the tunnel and on an open road are persistent characteristic humps observed in Hyundai cars running at 100 km/h, as shown in Fig. 5. 6. Noise Spectra on Flat, Ascending, and Descending Slopes The 1/3 octave band spectra of Equus running at 100 km/h on a flat, ascending, and descending slopes of almost 4 degrees for asphalt roads are compared in Fig. 6. The spectral profiles are virtually similar with regard to the SPL and spectral humps. 7. Noise Spectra of Gasoline and LPG Engines The 1/3 octave band sound pressure spectra of an Avante (1495 cc) with gasoline and liquid propane gas (LPG) engines running at 100 km/h on an asphalt road are presented in Fig. 7. The same spectral patterns were observed with nearly the same SPL and resonant peaks
Spectrum of Infrasound and Low-frequency Noise in Passenger Cars Sung Soo Jung et al. -2409- lower than Curve 1 except in the frequency ranges between 30 and 100 Hz, which may be a cause of complaint. This means that some passengers may have adverse feelings when they ride on cars. Curve 2 of the reference values, which was established in 2004 by Ministry of the Environment of Japan, was used to judge whether lowfrequency noise existed or not when complaints occurred [12]. Most measured noise levels were higher than Curve 2. Therefore, it is necessary to develop noise reduction methods in the low-frequency region. Fig. 8. 1/3 octave band spectra of Hyundai cars running at 100 km/h on an asphalt road. at 40 and 80 Hz, suggesting that the noise levels of the gasoline and the LPG engines are almost the same at the same car speed. 8. Sound Spectra of Various Cars of Hyundai The 1/3 octave band spectra of some typical Hyundai cars running at 100 km/h on an asphalt road are compared in Fig. 8. General spectral patterns of various Hyundai s cars are very similar, a nearly flat spectrum for 1 to 10 Hz, followed by a progressive decrease in the 1/3 octave band level with increasing frequency, regardless of engine capacity and mileage run. The smaller cars, the greater mileage runs, and rough roads are known to induce greater noise during running, and the sound pressure levels of small cars, such as Santa Fe (1991 cc, 78600 km) and Avante (1495 cc, 79000 km), are shown to be significantly higher than that of the larger Equus (3342 cc, 53000 km) in both inaudible and audible frequency ranges. In fact, the noise level of the small Santa Fe was the greatest among tested cars, suggesting that a diesel engine is much noisier than a gasoline or a LPG engine. It is also noticeable that the spectral hump at around 12.5 Hz appears persistently, regardless of car speed, and it is considered as one of the common features pertinent to Hyundai cars tested in the present work. IV. CONCLUSIONS The interior sound pressure levels of passenger cars of Hyundai that were standing idle and running at speeds of 80 to 110 km/h on flat asphalt and cement roads, including ascending and descending slopes, and tunnel roads were measured. Comparison of the narrow band noise spectrum of a car idling and the 1/3-octave band spectrum of the same car running at 100 km/h, but with the same engine speed, revealed that 1) An interior noise level of 45 db at infrasound and low frequencies of the car idling increased to 90 to 80 db for the car running at 100 km/h. The airborne rotating friction noise of tires/road may be the main source. 2) Humps at 80 and 40 Hz in the 1/3-octave band spectra of the running car corresponded to the first and the first sub-harmonics of the engine rpm resonances, respectively. There was virtually no difference in the spectral profiles between gasoline and diesel engines, ascending or descending slopes, and flat on sloping road. The interior infrasound level in the frequency range 10 Hz increased markedly with car speed, which can be a critical criterion for the assessment of ride quality. ACKNOWLEDGMENTS This subject is supported by the Ministry of Environment of the Republic of Korea as The Eco-technopia 21 project. 9. Psychological Effects of Low-frequency Noise REFERENCES The psychological effects of low-frequency noise in the interior of a passenger car were invetigated by comparing the existing criteria values. Curve 1 in Fig. 8 is the Japanese psychological limiting values (oppressive and vibratory sensation and discomfort feeling) for infra- and low-frequency noise [11]. The low-frequency noise levels measured in the interior of Hyundai s passenger cars are [1] D. T. Aspinall, MIRA Report No. 1966-2 (1996). [2] R. A. Hood and H. G. Leventhal, Acustica 25, 10 (1971). [3] W. Tempest and M. E. Bryan, Appl. Acoustics 5, 133 (1972). [4] M. J. Evans and W. Tempest, J. Sound Vib. 22, 19 (1972). [5] M. E. Bryan, Infrasound and low-frequency vibration (Academic Press, New York, 1976), p. 65.
-2410- Journal of the Korean Physical Society, Vol. 55, No. 6, December 2009 [6] H. E. Gierke and C. W. Nixon, Infrasound and lowfrequency vibration (Academic Press, New York, 1976), p. 115. [7] N. Broner, J. Sound Vib. 58, 483 (1978). [8] R. Guski, U. Felscher-Suhr, and R. Scheumer, J. Sound Vib. 223, 513 (1999). [9] H. G. Levanthall, Noise and Health 6, 59 (2004). [10] H. Onusic and V. S. Mizutani, J. Vehicle Design 37, 99 (2005). [11] H. Ochiai, Proceedings of Inter-noise 2001 (Hague, Netherland, 2001), p. 1495. [12] K. Kamigawara, J-I. Yue, T. Saito, and T. Hirano, Proceedings of low-frequency noise 2004 (Maastricht, Netherland, 2004), p. 157.