Meng-hua DUAN, Chen-guang LAI, Yong WANG, Jin-yang FENG, Wen-lin TAN

(1 Wind Tunnel Center of China Automobile Engineering Research Institute Co., Ltd., Chongqing 401122,China) (2 College of Vehicle Engineering, Chongqing University of Technology, Chongqing 400054, China ) (3 Institute of Fluid Science，Tohoku University，Sendai 980-8577， Japan)

Abstract: By using the separated eddy simulation (DES) and the fluid software as the research tool, the steady and transient simulation and analysis of the external flow field are carried out under the closed window condition in the rear-view mirror area of a vehicle. The pressure fluctuation on the body surface, which is used as the excitation to calculate the interior noise of the vehicle, is obtained. The coupling analysis of flow field and acoustic field around the rear-view mirror is realized by using acoustic software, and then the spectrum of sound pressure level is obtained. By analyzing the aeroacoustic characteristics of the internal and external fields of the basic model and the improved model, it is shown that the length of the rear-view mirror handle, the angle between the mirror handle and the mirror cover, and the shape of the mirror cover have great influence on the downstream flow field and the sound pressure level of the rear-view mirror. Compared with the basic model, the total sound pressure level on the outer surface of the front left-side windscreen of the improved model is reduced; the total sound pressure level in the inner field of the driver’s left ear is reduced by 6.41%, and the speech intelligibility is improved by 33.89%.

Key words: CFD, External noise, Internal noise, Acoustic coupling, Total sound pressure level, Speech clarity

With the increasing renewal of automobile technology and the increasing level of car ownership, the noise pollution of automobiles has motivated concern from consumers and the government. At the same time, the topic of noise reduction has attracted a large number of researchers. In recent decades, with the efforts of researchers, the application of new materials and the improvement of pavement conditions, mechanical noise, engine radiation noise, rotational noise and pavement noise have been well controlled [1- 4]. When the vehicle speed is higher than 80 km/h, the aerodynamic noise begins to occupy the main position of all noise [5]. When the vehicle speed exceeds 100 km/h, the aerodynamic noise becomes the dominant factor affecting the driving comfort [6-7].

The aerodynamic noise of automobiles is broadly divided into exterior field noise and interior field noise. The exterior field aerodynamic noise is proportional to the sixth power of the vehicle speed. That is, the higher the vehicle speed is, the higher the air velocity on the body surface is, the higher the amplitude of the fluctuating pressure on the vehicle surface is, and accordingly the higher the sound pressure level (SPL) corresponding to each frequency is [8]. The radiated noise transmitted to the interior of the vehicle by the fluctuating pressure acting on the panel wall is called the interior noise of the vehicle. Therefore, the fluctuating pressure on the automobile surface is the noise source of the automobile interior and exterior fields. In the stage of new vehicle development, it is necessary to optimize the body shape and bulge position in order to optimize the distribution of fluctuating pressure.

At present, the aerodynamic noise of automobiles is mainly studied by means of wind tunnel test and numerical calculation. Wind tunnel test is expensive and time-consuming, and it can only be carried out after the production of actual prototype. As a consequence, it is impossible to put forward constructive noise control scheme for automobile body design stage. By contrast, the numerical analysis is not affected by objective factors such as time, space and environment, but can also analyze the mechanism of noise generated by the interaction between fluid and solid surface. Therefore, numerical calculation has become an important research method of aerodynamic noise.

In this paper, the fluid and acoustic analysis approach are used to analyze the aerodynamic noise in the interior and exterior fields when the vehicle speed is up to 120 km/h. By comparing the spectrum information of sound pressure level between the basic model and the improved model, the mechanism of noise generation around the rear-view mirror is revealed, which provides the basis for the optimization and improvement to reduce the sound pressure level at the driver’s left ear and the surface sound pressure level outside the front window. Therefore, it has strong guiding significance and reference value for practical engineering application.

1 Aerodynamic noise theory

For the calculation of the external and interior noise near the rear-view mirror, the calculation domain of the whole vehicle is established and encrypted. Firstly, the RANS turbulence model is used for the steady state simulation. Then, the unsteady state simulation is carried out by using the separated eddy simulation (DES) turbulence model on the basis of the steady calculation to obtain the pressure fluctuation in the time domain. Afterwards, FTT transform and acoustic coupling from fluid flow field to sound field are carried out by Virtual Lab Acoustic to obtain the spectrum information of sound pressure level.

1.1 Separated Eddy Simulation (DES) theory

Detached-Eddy Simulation (DES) is an improved model based on the standard Spalart-Allmaras (S-A) equation, also known as the coupled turbulence model of large eddy simulation (LES) and Reynolds average N-S equation (RANS). DES model has the advantages of small amount of RANS calculation and high accuracy of large eddy simulation. In this paper, the Menter SST model of Reynolds average N-S equation model is used to control the transformation ofk-ωandk-εmodels under different conditions by switching parameters. The governing equation of Menter SST is as follows:

(1)

(2)

1.2 Aeroacoustic equation

For the flow field with moving fixed wall, Ffowcs Williams and Hawkins derive the following Ffowcs Williams-Hawkins equation from the Lighthill equation:

(3)

Where,pis turbulent pressure fluctuation,Tijis Lighthill tensor,pijis pressure on solid wall surface,δ(f) is Dirac delta function,a0,ρ0is sound velocity and fluid density, respectively. The first term on the right side of the equation represents the quadrupole sound source term caused by moving fluid outside the solid wall surface, the second term represents the dipole sound source caused by turbulent pressure fluctuation on the solid wall surface, and the third term represents the unipolar sound source caused by volume displacement effect. For high-speed vehicles in the air, the body surface can be regarded as a rigid body, and the volume displacement variation is almost zero, so the monopole sound source term can be neglected. As the ratio of quadrupole source intensity to dipole source intensity is proportional to the square of Mach number, the quadrupole source term can be neglected for vehicles with low Mach number. Then the aerodynamic noise outside the vehicle depends on the dipole source acting on the body surface by the pressure fluctuation. Therefore, the FW-H equation of vehicle flow field can be simplified from Eq. (3) to the following equation:

(4)

2 Establishing the boundary condition of dipole sound source on the body surface

Based on the model of a domestic car, the body data are as follows: lengthL=4 820 mm, widthW=1 800 mm, heightH=1 650 mm. The geometric model shown in Fig.1 (a). Fig.1 (b) is the original rear-view mirror model, and Fig.1 (c) is the improved rear-view mirror model. Compared with the original rear-view mirror model, Fig.1 (c) increases the minimum distance from the original rear-view mirror cover to the triangular window from 25 mm to 50 mm, changes the connection of the base of the mirror handle from obtuse angle to smooth transition, alters the angle between the inner surface of the mirror cover and the triangular cover plate from 10 ° to 0° and cancels the protrusion of the radar probe in the improved model. The wheel and chassis are simplified. The computational domains of length, width and height are 11L, 10Wand 5H, respectively. The volume meshes are divided and the meshes of rear-view mirror and side window are refined. In order to ensure the accuracy of the simulation results, grid independence validation was conducted for 40 million, 55 million and 70 million meshes, respectively. When the number of meshes was 55 million, the requirement of grid independence could be met.

Fig.1 Automobile geometry model, (a) vehicle model, (b) original rear-view mirror model, (c) improved rear-view mirror model

Firstly, the whole computational domain is computed steadily. When the fluid flow field is stable, the unsteady external flow field of a vehicle with a speed of 120 km/h is simulated and analyzed by using the separated eddy simulation (DES) method with the steady calculation results as the initial value. In unsteady simulation, the setting of time step is very important. It is related not only to the analysis of noise frequency, but also to the speed of the whole calculation. In the calculation, the time step is 2×10-5s. After the solver calculates 7 500 time steps, the pressure fluctuation data of the rear-view mirror and the left front window are collected, and the result data of 7 500 time steps are analysed. With the emphasis on the left front window noise and the interior noise, the interior and exterior acoustic models for generating acoustic boundary conditions are established as shown in Fig.2 (a) and Fig.2 (b). The monitoring points are located on the side windows and in the driver’s left ear. The number is shown in Fig.2. The pressure fluctuation data calculated by StarCCM+are imported into the acoustic model and converted into dipole source boundary conditions.

Fig.2 Acoustic cavity model of vehicle interior and exterior field and monitors, (a) closed acoustic model including front window, A-pillar and rear-view mirror area and three monitoring points, (b) acoustic model of vehicle interior field noise and driver’s left ear monitoring point, (c) monitoring point of the vehicle interior

3 Calculation results and analysis

3.1 Flow field analysis

For the aerodynamic noise around the rear-view mirror, the fluctuating pressure on the front windowpane surface is the main noise source, while the wake vortex region behind the rear-view mirror is the source of the fluctuating pressure. The flow field in the whole fluid field can be obtained by steady calculation. The height of cross section is 0.6Hin theZaxis. Fig.3 and Fig.4 are pressure distribution and turbulent kinetic energy distribution contours on the cross section, respectively. From the comparison of Fig.3(a) and Fig.3(b), it can be found that the area of the high pressure zone on the windward side of the rear-view mirror is larger than that of the improved model, and the pressure fluctuation between the rear-view mirror and the side window is larger than that of the improved model, because the length of the improved model handle increases, and the negative pressure area from the rear-view mirror cover to the side window section is easier to be filled with the surrounding air.

Fig.3 Pressure distribution contours on cross section around rear-view mirror. (a) Pressure Distribution on Section of Foundation Model (b) Pressure Distribution on Section of Improved Model

Turbulence kinetic energy is a measure of turbulence intensity. The magnitude of turbulence kinetic energy can reflect the intensity of turbulence and the intensity of pressure fluctuation in this region. By comparing the turbulent kinetic energy distribution contours of Fig.4, it can be found that both the improved model and the basic model have strong turbulent distribution in the wake region of the rear-view mirror; the turbulent region in Fig.4 (a) is close to the side window, and strong turbulence hits the side window, while the turbulent region in Fig.4 (b) is a certain distance from the side window, and the turbulent disturbance caused by the rear-view mirror does not flow through the side window. The turbulent kinetic energy distribution of the two models coincides with the pressure distribution analysis.

3.2 Aerodynamic acoustic field analysis

3.2.1 External field noise analysis

The surface turbulent pressure fluctuation information of rear-view mirror, side window, A-pillar and triangular cover plate is transferred through data and transformed into dipole sound source by fast Fourier transform to calculate aerodynamic noise. Since the vehicle speed is less than Mach 0.3, the influence of quadrupole source could be neglected in acoustic calculation. Table 1 shows the distribution of sound pressure level on the outside surface of the side window before and after the improvement by calculating the acoustic response. Through comparative analysis, the middle area of the improved side window surface has a strong sound pressure level distribution, which is close to the driver’s left ear; after improvement, the area of high sound pressure level is concentrated in the front area of the side window in three frequency bands, mainly because of the improvement of the structure of the rear-view mirror, resulting in better flow field characteristics, and there is no strong turbulence impact on the side window downstream of the rear-view mirror.

Fig.4 Turbulent kinetic energy distribution contours. (a) Turbulence on the cross section of the basic model (b) Turbulent kinetic energy distribution on the cross section of the improved model

Table 1 Comparison of sound pressure level distribution contours at 600, 1 000 and 1 600 Hz on the outside surface of side window before and after improvement

After obtaining the results of the acoustic response of the side window surface, the frequency of the sound pressure level at each monitoring point is calculated by the sound pressure frequency response function. The calculation results are shown in Fig.5 and the layout of the monitoring points is shown in Fig.2 (a).

Fig.5 Spectrum curve of sound pressure level at monitoring point

Fig.5 shows that the sound pressure level of monitoring point 1 is slightly larger than that before improvement in the frequency band of 100-500 Hz, and slightly larger than that before improvement in a few frequency bands of 4 000-6 000 Hz. Overall, there is little difference between the sound pressure level before and after improvement in monitoring point 1, and the sound pressure level in the whole frequency band after improvement in 2, 3 monitoring points is 10 dB smaller than that before improvement. Through the analysis of turbulence intensity downstream of rear-view mirror, the reason why the sound pressure level decreases at monitoring points 2 and 3 after the improvement of the model is explained.

3.2.2 Interior noise analysis

After obtaining the sound pressure outside the vehicle, the sound pressure is loaded on the window glass as an excitation. Through direct acoustic-vibration coupling calculation, the frequency response of the sound pressure inside the vehicle can be obtained. In this paper, the sound absorption coefficients of the interiors are not set separately, but the equivalent sound absorption coefficients of each frequency band are calculated by reverberation time, which is loaded into the imaginary part of the air sound speed to obtain the accurate sound pressure level distribution in the interior of the vehicle.

Fig.6 Improvement of 1/3 octave sound pressure level curve in vehicle

Fig.6 is the contrast curve of 1/3 octave sound pressure level in the driver’s left ear (Fig.2(c)) before and after the improvement of rear-view mirror. It can be seen from the figure that the maximum sound pressure level of 1/3 octave is reduced by 3.18 dB, and the sound pressure level is also reduced in the middle and high frequency bands. According to the national standard of speech intelligibility, the medium and high frequency bands contribute a lot to speech intelligibility. Therefore, in Table 2, the speech intelligibility is improved by 33.89% after the improvement, which is due to the improvement of the sound pressure level in the middle and high frequency bands. The total sound pressure level is calculated by superimposing the sound energy of each frequency with the formulaSPL=20*lg(P/P0), where,Pis the sound pressure,P0is the reference pressure. After improvement, the total sound pressure level decreases by 4.29 dB.

Table 2 Comparison of Total Sound Pressure Level and Speech Intelligence before and after Improvement

4 Conclusion

Through the combination of DES transient and acoustic calculation, the three-dimensional transient flow field and aerodynamic noise inside and outside the vehicle are simulated and analyzed. After the improvement of the geometry and layout of the rear-view mirror, the flow field near the rear-view mirror is improved, the wind noise inside and outside the rear-view mirror is reduced, and the speech intelligibility inside the vehicle is improved. It is revealed that the flow field characteristics can be improved, some aerodynamic noise can be reduced and design in the rear-view mirror area, which provides guidance for the design of rear-view mirror.