# Prediction interior noise excitation force the powertrain based on hybrid transfer path analysis_图文

International Journal of Automotive Technology, Vol. 9, No. 5, pp. 577?583 (2008) DOI 10.1007/s12239?008?0068?8

Copyright ? 2008 KSAE 1229?9138/2008/042?08

PREDICTION OF INTERIOR NOISE BY EXCITATION FORCE OF THE POWERTRAIN BASED ON HYBRID TRANSFER PATH ANALYSIS

Hybrid TPA与传统TPA?方法的区别 S. J. KIM and S. K. LEE 在于激励的来源，传统 TPA 的激励数 Department of Mechanical Engineering, Inha University, Incheon 402-751, Korea 据是试验获得的，Hyper TPA的激励 (Received 20 February 2008; Revised 30 May 2008) 数据是计算出来的。

ABSTRACT?In the early design stage of a vehicle, simulation of interior noise is useful for assessment and enhancement of the noise, vibration and harshness (NVH) performance. Traditional transfer path analysis (TPA) technology cannot simulate interior noise since it uses an experimental method. In order to solve this problem, hybrid TPA is employed in this paper. Hybrid TPA uses simulated excitation force as the input force, which excites the flexible body of a car at the mount points, while traditional TPA uses the measured force. This simulated force is obtained by numerical analysis of the finite element (FE) model of a powertrain. Interior noise is predicted by multiplying the simulated force by the vibro-acoustic transfer function (VATF) of the vehicle. The VATF is the acoustic response in the compartment of a car to the input force at the mount point of the powertrain in the flexible car body. The trend of the predicted interior noise based on the hybrid TPA corresponds very well to the measured interior noise, with some difference due to not only experimental error and simulation error, but also the effect of the airborne path. KEY WORDS : Powertrain, Force prediction, FEM, Structure-borne noise, Vibro-acoustic

*

1. INTRODUCTION

Noise vibration harshness (NVH) technology in automobile engineering has become an important performance target in the search for vehicle comfort. In a vehicle, there are many noise and vibration sources that influence NVH performance, such as the powertrain, tires, wind, car body, suspension, etc. Among these sources, it is known that the powertrain is the dominant contributor to interior noise (Lee et al., 1994). Therefore, many researchers have tried to develop a simulation method to predict interior noise caused by a powertrain (Seki et al., 2001; Kim et al., 2007). Interior noise due to a powertrain is induced by two transfer paths: airborne (noise) and structure-borne (vibration). In order to predict interior noise due to the structureborne transfer path, traditional transfer path analysis (TPA) has been used based on an experimental method (Wyckaert and Auweraer 1995; Lee et al., 2000). Although it is a useful tool for source identification, it is not convenient for modification of the powertrain structure for reducing interior noise since it uses an experimental method. This paper presents a hybrid method for the prediction of interior noise based on hybrid TPA. In order to solve the interior noise problem, the numerical model of a car body was used and the excitation force was measured (Auweraer et al., 2007). In an ooother paper, we discussed the use of the numerical load combined with experimental FRSs (Marco et al., *Corresponding author. e-mail: sangkwon@inha.ac.kr

577

K

2006). For calculation of the numerical load, we used the ADAMS model instead of the FE model of a real engine. The moving part components of the engine were all rigid and the powertrain was the virtual FE model built in ADAMS. Therefore, their results are not pratical and do not simulate a real situation. However, the numerical load for a real engine is calculated based on the FE-model of a real powertrain by other authors (Lee et al., 2006) and the vibro-acoustic characteristic for a real car is measured for hybrid TPA. The simulation of the exciting force is validated by the measured exciting force. The measured exciting force is estimated by multiplying the dynamic complex stiffness of the mount rubber by the displacements measured on the mount bracket before and after the isolation of the mounting system of the powertrain. The dynamic complex stiffness of the rubber isolator is measured by the elastomer testing system with an electric actuator and isothermal reservoir. We also predicted the interior noise based on hybrid TPA. The simulated exciting force is used in the prediction of this interior noise. In order to predict the interior noise based on hybrid TPA, the VATF of the vehicle is also necessary and is measured by using the vibro-acoustic reciprocity method (Kim and Ih, 1993; Ko et al., 2006). The predicted interior noise is compared with measured interior noise. The trend of the predicted interior noise corresponds to the direct measured noise well, although there are some differences in the absolute magnitude. 计算与实验趋势?一样仅有微?小误差

F=K*S

578

S. J. KIM and S. K. LEE

2. THEORY OF HYBRID TRANSFER PATH ANALYSIS

The traditional TPA method is the experimental method to predict interior noise (Wyckaert and Auweraer, 1995). This method is useful for the identification of noise sources along the vibration transfer path. However, after identification of the noise source, if a design modification of the powertrain structure is necessary for the reduction of interior noise a simulation tool for the prediction of the interior noise is required. In this case, hybrid TPA is one of the best solutions. The prediction of interior noise inside a car based on traditional TPA is expressed as follows: pj ( ω ) = ∑ p ( ω ) i

i jφ pj( ω ) = ∑ f i ( ω ) i fi -------------fj ( ω ) i j ( φ P ( ω ) – φf ( ω ) )

i

j φ ( ω ) fi

+ ∑ qk ( ω )

k

j φ qi

pj ( ω ) ------------qk ( ω )

j ( φ P ( ω ) – φq ( ω ) )

k

(1)

where pj is the interior sound pressure at point i inside a car, fi is the excitation force at the i-th path, qk is the air-borne source at the k-th path and φ is the phase corresponding to each source and response. The first term of equation (1) is related to the structure-borne path and the second term is related to the airborne path. Interior noise based on the traditional TPA for the structure-borne path is predicted by using the first term in equation (1). Therefore, equation (1) can be rewritten by neglecting the second term as follows: p j ( ω ) = ∑ p i ( ω ) = ∑ f i ( ω ) × H ij ( ω )

i=1 i=1 n n

Figure 1. Diagram for the powertrain mounting system of the test vehicle. paths of powertrain vibration for interior noise among many transfer paths in the vehicle. These mounts play the role of supporting the weight of the powertrain and absorbing vibrational energy. These mounts are composed of the isolation rubber and bracket. The combustion force due to the firing of the test engine excites the piston and its

(2)

where p i ( ω ) is the complex sound pressure due to the i-th path in the frequency domain. In equation (2) the vibroacoustic transfer function of a vehicle, H ij ( ω ) is obtained by using the experimental method based on the theory of the vibro-acoustic reciprocity. The excitation force f i ( ω ) is calculated by multiplying the dynamic stiffness by the displacement difference based on traditional TPA. The mathematical expression for excitation force is given by, fi ( ω ) = k ( ω ) × Δ x ( ω ) = ki ( ω ) [ x 1 ( ω ) – x2 ( ω ) ]

*

(3)

where k i ( ω ) is the complex dynamic stiffness measured by the elastomer testing system with electric actuator. In equation (3), the displacements x 1 ( ω ) and x 2 ( ω ) are measured on the mount bracket before and after isolation, respectively, while a car is being driven. In general, the powertrain is installed on the car body by using a mounting system as shown in Figure 1. This mounting system consists of four mounts: engine (E/G) mount, front-roll (F/R) mount, rear-roll (R/R) mount, and transmission (T/M) mount. The mounting system is connected to the sub frame or the suspension frame of a vehicle. Therefore, these mounts are the major transfer

Figure 2. Diagram of the hybrid transfer path analysis for using the FE model of a powertrain.

PREDICTION OF INTERIOR NOISE BY EXCITATION FORCE OF THE POWERTRAIN BASED

579

dynamic force is transferred to these brackets through the cylinder block. This dynamic force also excites the car body through the isolation of the mount system. The dynamic force exciting the car body can be obtained by two methods. One is the experimental method based on the traditional TPA and its mathematical form is given by equation (3). In order to use equation (3), the operational vibration should be measured on the mount bracket after and before the isolation of the mount system of the powertrain while a car is being driven. Traditional TPA uses the exciting force obtained by this experimental method. Hybrid TPA uses the simulated exciting force instead of the measured exciting force. The simulated exciting force is calculated by numerical analysis for the FE model of a powertrain (Kim et al., 2008) as shown in Figure 2. The advantage of hybrid TPA is the possibility of modifying the powertrain structure because it uses the FE model of a powertrain. The model includes engine mounts and their boundry conditions are given by measuring the mobility at the mount point of the car body.

Figure 4. Complex dynamic stiffness measured by elastomer testing system: (a) E/G mount; (b) F/R mount; (c) R/R mount; (d) T/M mount.

3. MEASUREMENT OF EXCITING FORCE OF THE POWERTRAIN

In order to calculate the exciting force of the powertrain through each isolator by using equation (3), the complex stiffness of each isolator and the operational displacement are required. The complex stiffness of each isolator is measured by using the elastomer-testing system as shown in Figure 3. The complex stiffness is related to the complex force, which is provided by the electric shaker supplier and acts on the steel block encompassing the powertrain mount. The complex displacement, which is measured by the accelerometers. Figure 4 shows the magnitude and phase of dynamic stiffness for powertrain mounts. The dynamic stiffness for powertrain mounts is measured in three rectangular directions, considering the direction of the vehicle. The operational displacement xi is obtained by doubly integrating the accelerations measured on both sides of each isolator while the test car is being driven on a chassis

K

Figure 5. Photograph for measurement points of accelerations at the powertrain mounts. dynamometer in an anechoic chamber. Figure 5 shows the accelerometers (B&K model 4507B004) attached on the bracket of each mount. The accelerometers measure the operational accelerations at the bracket before and after the isolators. The complex dynamic stiffness as shown in Figure 4 and operational displacement are used for the estimation of the measured excitation force by using equation (3). Figure 6 shows the direct comparison between the measured exciting forces and the simulated exciting force, which is obtained by numerical analysis for the FE model of a powertrain based on the hybrid TPA (Kim et al., 2008). In previous work, the simulated excitation force was validated indirectly by comparing the measured displacements with the simulated displacement at the mount brackets (Kim et al., 2008). According to these results, both forces increase with the increase of the rotation speed of the

S

Figure 3. Elastomer testing systems for measurement of dynamic stiffness.

580

S. J. KIM and S. K. LEE

Figure 6. Comparison between the measured excitation force and the simulated excitation force: (a) 2nd order component; (b) 4th order component; (c) 6th order component. crankshaft. There is some difference in the absolute level, however. This difference is due to the measurement error for the measured force and the numerical error for the numerical force. The force at the engine mount is higher than that at the other mounts. The 2nd order force component of the rotation speed of the crankshaft is higher than the 4th order and 6th order force component. Although these results are not sufficiently satisfactory, it is still useful to simulate and predict the characteristics of the structure-borne path of the powertrain vibration. In order to reduce the numerical and experimental error, more trials of the research are required in future work. These exciting forces are used for the prediction of the structure-borne noise. where pj is the sound pressure at the j-th point inside the car and fi is the excitation force at the i-th point of the flexible car body. The VATF for a test car provides information for the acoustic response of the cavity excited by the excitation force of the powertrain. Measuring the VATF for a test car using equation (4) is a long and tedious process because there are many transfer paths. In order to measure the VATF for a test car with high accuracy in a short time, vibro-acoustic reciprocity is employed (Kim and Ih, 1993) in this paper. According to the theory of vibro-acoustic reciprocity (Ko et al., 2006), the VATF for a test car is also expressed as follows: vi ( ω ) H ij = ------------qj ( ω ) (5)

4. VIBRO-ACOUSTIC TRANSFER FUNCTION

The vibro-acoustic transfer function for a car is the acoustic response in the compartment of the car to the force exciting the car body at the transfer path and is defined by: pj ( ω ) H ij ( ω ) = ------------fi ( ω ) (4)

where q j ( ω ) is the volume source at the j-th point inside a car and v i ( ω ) is the vibration response at the i-th point of the flexible car body. In order to obtain the VATF of a test car by using equation (5), the volume source made by the LMS company is installed in the front seat of the test car as shown in Figure 7. In order to confirm the theory of vibro-acoustic reciprocity, first, the input force is excited at the engine mounts

PREDICTION OF INTERIOR NOISE BY EXCITATION FORCE OF THE POWERTRAIN BASED

581

Figure 7. Setup of the equipment for the measurement of the vibro-acoustic transfer function in a test vehicle. Figure 9. Measured vibro-acoustic transfer function based on reciprocity theory.

Figure 8. Measured vibro-acoustic transfer function based on the reciprocity theory. and the sound pressure at the front seat of the test car is measured. The VATF for the test car is calculated by using equation (4). Secondly, the volume velocity excites the cavity at the front seat and the vibration at the engine mount point is measured. The VATF for the test car is calculated by using equation (5). Figure 8 shows the VATF for the test car obtained by both methods, which are compared. The solid line represents the results obtained by using equation (5), and the dotted line represents the results obtained by using equation (4). The two results are very close, and thus the theory of vibro-acoustic reciprocity is verified. Finally, the VATF for each mount point of the test vehicle is measured. Figure 9 shows the VATF for each mount obtained by using equation (5) in three directions. These VATFs for the each mount point of the test vehicle are used for the prediction of the interior noise. Figure 10. Image map for the interior noise. measured .inside a test vehicle. from 1000 rpm to 4000 rpm. The specification of the test vehicle is listed in Table 1. The test car is a sport utility vehicle with an in-line 4 cylinder engine. Figure 10 shows the image map for the interior noise measured at the front seat. According to the results, the 2nd, 4th and 6th orders of the revolution of the crank shaft are dominant. The interior noise caused by the structure-borne path of the powertrain vibration is predicted by hybrid TPA, and their results are compared with the one obtained by traditional TPA as shown in Figure 11. According to these results, the predicted noise based on hybrid TPA corresponds to that based on traditional TPA, although there is some difference. Figure 11 shows the contribution of transfer paths to the interior noise. This is the contribution of only the structure-borne path of the powertrain vibration. This contribution is summed and its result is compared with the interior measured inside a car as shown in Figure 12. Figure 12 indicates the difference between the predicted noise and the measured noise. This difference at low speed is due to the airborne

5. PREDICTION OF INTERIOR NOISE BASED ON THE HYBRID TPA

The interior noise for the test vehicle is measured at the front seat while the car is driven at a crankshaft rotation

582

S. J. KIM and S. K. LEE

Figure 11. Prediction of the interior noise based on the traditional TPA and hybrid TPA for the harmonic components of the rotating speed of the crankshaft: (a) 2nd order component; (b) 4th order component; (c) 6th order component. Table 1. Specification of the powertrain. Powertrain Displacement Number of cylinders Type of fuel Balance shaft Transmission Vehicle type Specification 2.2 L 4 Diesel Lanchester type Automatic SUV

Figure 12. Comparison between the direct measured interior noise and the predicted interior noise based on the hybrid TPA: (a) overall noise; (b) 2nd order component; (c) 4th order component; (d) 6th order component. path of the powertrain. It may be from the contribution of the intake and exhaust system. The difference at high speed

is due to the airborne path of the tires or wind. However, hybrid TPA is very useful for the prediction of interior noise based on simulation technology because it is possible to estimate the trend of the structure-borne noise and is applicable to the optimization of the powertrain structure. Hybrid TPA is also applicable to the prediction of sound quality (Lee and Chae, 2004; Lee et al., 2005) by using the sound quality index.

6. CONCLUSION

In the paper, the hybrid TPA is developed for the prediction

PREDICTION OF INTERIOR NOISE BY EXCITATION FORCE OF THE POWERTRAIN BASED

583

of the interior noise caused by the structure-borne path of the powertrain. This technology is useful for design modification of the powertrain in the early development stages of the car to tune the target noise. In this technology, the FE model of the powertrain is used for the simulation of the excitation force, which is the input force of the flexible car body at the mount points of the powertrain. This simulated excitation force is compared with the measured force. The predicted interior noise is also compared with the interior noise measured inside a car. The results show the trend of the simulated excitation force and the interior noise corresponds to the measured force and interior noise well, although there is some difference between the absolute levels. Therefore, hybrid TPA is very useful for the prediction of interior noise based on the simulation technology and it is applicable to the optimization of the powertrain structure.

ACKNOWLEDGEMENT?This work was partly supported by Automobile Component Base Technology Project No. 10023237 in Korea.

REFERENCES

Auweraer, H. V., Mas, P., Dom, S., Vecchio, A., Janssens, K. and Ponerrle, P. V. (2007). Transfer path analysis in critical path of vehicle refinement: the role of fast, hybrid and operational path analysis. SAE Paper No. 2007-01-2352. Kim, B. G. and Ih, J. G. (1993). In-Situ estimation of an acoustic source in an enclosure and prediction of interior noise by using the principle of vibroacoustic reciprocity. J. Acoustical Society of America 93, 5, 2726?2731. Kim, S. J., Kim, S. G., Oh, K. S. and Lee, S. K. (2008). Excitation force analysis of a powertrain based on CAE technology. Int. J. Automotive Technology, Submitted. Kim, S. J., Lee, J. Y. and Lee, S. K. (2007). Noise refinement of a vehicle by reduction of the axle gear whine noise based on structural modification using FEM and BEM. Int. J. Automotive Technology 8, 5, 605614.

Ko, K. H., Kook, H. S. and Heo, S. J. (2006). New technique in the use of vibro-acoustical reciproicity with application to the noise transfer function measurement. Int. J. Automotive Technology 7, 2, 173?177. Lee, J. H., Lee, S. K. and Kim, S. J. (2006). Anslysis of excitation forces for the prediction of the vehicle interior noise by the powertrain. Trans. KSNVE. J. 16, 12, 1244? 1251. Lee, S. K., Hwang, W. S., Kim, J. H., Woo, J. H., Lee, S. H. and Lee, H. J. (2000). Improvement of sound quality of vehicle through reduction of interior noise using noise transfer path analysis and running modal analysis. KSNVE. J. 10, 5, 801?806. Lee, S. K. and Chae, H. C. (2004). The application of artificial neural networks to the characterization of interior noise booming in passenger cars. Proc. Instn. Mech. Engrs.: Part D 218, 1, 33?42. Lee, S. K., Kim, B. S. and Park, D. C. (2005). Objective evaluation of the rumbling sound in passenger cars based on an artificial neural network. Proc. Instn. Mech. Engrs.: Part D 219, 4, 457?469. Lee, S. K., Yeo, S. D., Kim, B. J. and Rho, I. H. (1994). Weight reduction and noise refinement of hyundai 1.5 liter powertrain, SAE Paper No. 940995. March, J., Poggi, M., Maunder, R., McGregor, N., Powell, N. and Strong, G. (2006). The application of flexible multi-body dynamic model and noise transfer path analysis to optimize vehicle sound quality. FISITA 2006. F2006D206. Sakai, T. and Sakamoto, A. (2003). Improvement of engine noise for the 2003 accord using hybrid CAE technology. SAE Paper No. 2003-01-1427. Seki, Y., Suzuki, T., Tsukahara, M. and Takahashi, Y. (2001). How to predict powertrain vibration at the engine mounting points under running conditions. SAE Paper No. 2001-01-1592. Wyckaert, K. and Auweraer, H. V. (1995). Operational analysis, transfer path analysis, modal analysis: Tools to understand road noise problem in cars. SAE Paper No. 951752.