Technical issues of nearfield measurements inside a car for improving acoustic comfort Possibilités techniques d’antennerie concernant l’amélioration du confort en habitacle Daniel Vaucher de la Croix - Patrick Chevret Metravib RDS - Limonest Abstract A continuously growing demand comes from the automotive industry for optimization of materials and sound insulating products implementation inside the car, so as to propose the best acoustic performance at reduced costs. The acoustical holography system LORHA developed by METRAVIB RDS provides part of the solution to such a demand. Its demonstrated capability of measuring the acoustic field inside a vehicle makes it an advanced tool for performing extensive studies of the acoustic transparency of car openings as well as wind tunnel measurements. LORHA allows now for : ü ü ü ü
detailed localization of noise sources or acoustic weakness points inside the vehicle, knowledge of the acoustic energy distribution on elementary surfaces (such as doors, windscreen, roof, sealing system, …), reconstruction of the energy radiated by elementary surfaces in order to predict the acoustic pressure at the driver’s and passengers’ ears, estimation of the acoustic incidence of local modifications on components of the tested car.
The proposed technique will allow to apply new acoustically driven design methods for a complete understanding and control of all components and systems involved in the car performance. Résumé Une demande forte de la part des industriels de l’automobile est formulée actuellement pour disposer d’un outil expérimental permettant d’optimiser l’utilisation de matériaux et de produits insonorisants ou isolants, dans le but de proposer la meilleure performance acoustique au meilleur coût. Le système d’imagerie acoustique par holographie de champ proche LORHA proposé par METRAVIB RDS apporte des éléments de réponse à cette attente. Sa capacité de mesure avérée en intérieur habitacle de véhicule automobile lui confère un réel avantage pour la conduite des études de transparence d’ouvrants aussi bien en chambre réverbérante qu’en soufflerie. En ce sens, LORHA permet aujourd’hui : • • • •
de localiser très précisément les sources ou fuites acoustiques en quasi contact avec l’objet mesuré par cartographie 3D, de connaître les répartitions d’énergie acoustique sur plusieurs surfaces ou sous-surfaces élémentaires (portes, pare-brise, pavillon, joints d’étanchéité, …), de reconstruire l’énergie rayonnée par plusieurs surfaces élémentaires pour le calcul de la pression acoustique en certains points de l’habitacle (oreilles conducteur et passagers), de recalculer le niveau de pression acoustique à l’intérieur de l’habitacle consécutif à ces changements. On obtient alors l’effet d’un changement de composant habitacle sur le confort acoustique des occupants.
La technique proposée doit contribuer à la mise au point des outils d’aide à la conception de composants de véhicule dans un contexte d’optimisation des performances acoustiques.
1. INTRODUCTION Acoustical comfort inside vehicles tends to become a dimensioning element that manufacturers integrate even sooner into their design approach of a new model. Knowing the diversity of possible transfer paths of various noise sources into the interior of the vehicle, they also intend to share this preoccupation with their major equipment manufacturers. The latter ones have thus to think about the best possible design of their products, integrating the acoustical constraints translating the car manufacturers’ expectations into specifications. This paper aims at describing the approach adopted by METRAVIB RDS regarding this matter, with a special attention focused on both methodology and operational aspects of the inside vehicle measurements. Typical results obtained from advanced post-processing tools highlight the relevance of the proposed analysis for :
Astelab 2003
28 -
1
Technical issues of nearfield measurements inside a car for improving acoustic comfort 1. the localization of acoustic sources (mapping of relative contributions from different panel components, i.e. sealing system), 2. the prediction of the acoustic pressure at various locations inside the car (more specifically at driver’s and passengers’ ears), 3. Numerical masking for component analysis. In a first section of the paper, we present both the theoretical fundaments of the holography technique and the global methodology involved during inside vehicle measurement. Second, we show some typical results (on the base of comparisons with classical measurement techniques such as accelerometers, velocimeters, etc.)
2. 2.1
METHOD PROPOSED - WORK HYPOTHESIS – ADVANTAGES Theoretical formulation of the Nearfield Acoustical Holography
The fundamentals of nearfield acoustical holography may be found in many scientific publications. Major papers need to be kept in mind by the reader [1-8]. Among them, reference [1] is certainly the most cited in the literature. The basic idea is to take advantage of the complexity and richness of the source nearfield information (radiated and flexural field). This is performed by realizing a set of acoustical measurement in a plane which should be parallel and cover the radiating source (figure 1).
Figure 1 : Acoustical plane holography - coordinates system ( d is the separation distance between the source plane (white) and the measurement one (gray))
The different theoretical steps involved in the holography processing are briefly presented in the following. First, the measured acoustical field (on a plane regular grid) is decomposed on a set of 2D plane waves (on the base of a simple spatial Fourier transform). The dual space to the coordinates space is composed the orthogonal wave number components
k ⊥ (k x , k y ) = k z of the plane wave spectra. The pressure spectra in the wave number space is obtained from the following
expression:
(
)
pˆ k x , k y ,0, ω =
+∞
∫ p (x, y,0,ω)exp (ik −∞
x x + ik y y
) dxdy = TF ( p (x, y ,0,ω)) ,
Second, the obtained plane wave spectra is back-propagated from the measurement plane to the source plane. This operation r r requires the knowledge of the propagation Green function (noted G r / r0 , k z ). Without any filtering operation, the
(
)
acoustic pressure field is recomposed on the source plane with the help of the inverse Fourier transform, according to the expression :
Astelab 2003
28 - 2
Technical issues of nearfield measurements inside a car for improving acoustic comfort
((
)
)
r r p ( x, y , d , ω ) = TF −1 pˆ k x , k y , d , ω × G −1 (r / r0 , k z )
=
[
]
.
r r p (x , y ,0, ω ) ⊗ TF −1 G −1 (r / r0 , k z )
. Formally speaking, there is no difficulty in applying this succession of operations. In practice, there is a need to apply relevant filters during the back propagation phase. This is due to the very complex structure of the near acoustical field. To simplify, the nearfield can be viewed as the addition of a purely propagative part (active one which can be recovered in the far field of the structure) and an evanescent part (reactive one due to flexural wave propagation on the structure) which amplitude decreases exponentially with the separation distance. Back-propagating the measured pressure field leads to amplifying both the evanescent components and, often dramatically, the ambient noise. Among all the filtering possibilities, the Wiener filter (optimal filter) is used to remove the uncorrelated source of noise from the signal of interest. It consists in replacing the inverse Green function on the previous equation by the following operator:
(
)
Gw k x , k y , dz =
(
(
G ∗ k x , k y , dz
)
)
G 2 k x , k y , dz + N 2 S 2
.
N 2 et S 2 are the autocorrelation spectra of respectively the noise and signal on the source plane. Another type of filter can be advantageously added to the optimal one in order to reduce the extra amplification of evanescent component, especially in the low frequency domain: the Veronesi filter which acts like a low pass filter and whose expression and application may be find in the following references [2].
2.2.
Experimental methodology
Nearfield Acoustical Holography (NAH) makes it possible to obtain noise maps of a test structure (heat engine, submarine, fan, car openings, …) from nearfield pressure measurements. Another specificity of this technique lies in the possibility to get the acoustical radiation at a given distance from the sources. This is performed thanks to the use of the Helmholtz Integral Equation. For more than 5 years, METRAVIB RDS has transposed this technique to the specific application of inside vehicle measurements which first objective is to access to the acoustical radiation of system components (through a measurement series made inside the vehicle, with a physical, real or artificial excitation situated on the outside). The car and equipment manufacturers’ challenging request to obtain the necessary data for such a precise analysis lies in the operational use of the measurement means which has to meet very reduced handling time requirements. Today, the time needed for performing a complete information recording on a typical car (the lateral sides, the windscreen and the roof), including a first analysis in the whole measurement series, is reduced to about half a day or less. This is achieved thanks to : 1.
the use of a compact measurement means, light and easy to handle, particularly adapted to the specific situation inside the vehicle. In this case, it is a square hand-held array composed of 64 microphones spaced out of 2.5 cm (figure 2a),
2.
the recording of the geometrical information relative to the exact positioning of the antenna, combined with an automatic release of the measurement once the antenna location has been recorded (figure 2b),
3.
the implementation of a diffuse interpolation technique which eases the operator from measurement requirements according to regular geometry’s (figure 2c).
Astelab 2003
28 - 3
Technical issues of nearfield measurements inside a car for improving acoustic comfort
(a) Inside vehicle measurements using a dedicated square antenna
(b) The positioning system
(c) Example of a set of position of the acoustical racket
(d) Regular grid used for the holographic processing
Figure 2: Measurement and positioning system pictures and example of post-processing result
3. EXAMPLE OF APPLICATION One of the main advantage of the holographic approach (aside from the nearfield characteristic of the measurement) lies in the full-coherence of the analysis. At any time of the treatment, the acoustic field is equivalent to a set of plane waves which amplitude and phase are perfectly known. This allows a large type of post-processing (both on the entire surface of measurement as well as on elementary surfaces, which are not possible with standard measurement approaches) such as the computation of the attenuation index of specific components (the sealing system of the car openings for instance, which is a major actor to acoustic transfers and noise distribution inside the vehicle) or the predictions of acoustic levels at occupant’s locations inside the vehicle. In the following, we present the results of post-processed holographic measurement performed both on the lateral face and inside a medium size car. Comparisons with other type of measurement approach (acoustics, accelerometer and velocimeter based measurements) are made in order to validate the holographic approach.
3.1. Localization, quantification and ranking of the acoustic sources A series of measurement is conducted on a car of the medium category placed in a reverberant room. A wide band white noise is generated in the chamber (outside the car) and the acquisition is performed in the car, on a vertical plane, at 10 cm from one lateral side. The other faces (opposite lateral face, windscreen, roof) are masked in order to minimize the acoustic energy coming elsewhere from the face of interest. Overall measurement time to fully cover one side of the car is about 1/2 hour or even less. Astelab 2003
28 - 4
Technical issues of nearfield measurements inside a car for improving acoustic comfort The sound pressure field on the lateral side is calculated according to the theoretical and operational elements described in the previous section. Results may be presented in terms of pressure and (as an advantage of the acoustical holography approach) velocity or intensity, on a narrow band, third octave band, etc. 3.1.1. Mapping of the acoustical field on a lateral panel The maps presented on figure 3 below show the relevance of the holographic processing for the source localization and ranking. Clearly the back propagation operation enables to identify the precise regions of weakness on the lateral face (figure 3b –1000 Hz and 3c –4000 Hz), on the contrary to the raw data measured 10 cm away from the radiating plane (figure 3a and 3b). For instance, it appears for both frequencies that the sealing system at the top front of the door needs to be improved as well as at the bottom right of the front lateral windscreen.
(a) Measured pressure field (10cm from surface) 1000 Hz
(b) Back – propagated pressure field 1000 Hz
(c) Measured pressure field (10cm from surface) 4000 Hz
(d) Back – propagated pressure field 4000 Hz
Figure 3: Example of source localization on a lateral face of the car 3.1.2. Prediction on the lateral face of the car On a quantitative point of view, many confrontations with standard measurement techniques (accelerometer and velocimeter based) have been performed. Figure 4 below presents typical results of these comparisons. The curves show the velocity spectrum on the front and rear windscreens measured with an accelerometer (in black), with a laser velocimeter (in blue) and the one predicted by holography processing. In both cases, the accelerometer measurement and the holographic prediction are very similar over the whole frequency range. The velocimeter measurements are open to criticism, especially on the rear lateral screen. Indeed, such a technique is not well adapted to measurement on a glassy structure as it requires an homogeneous painting of the surface.
Astelab 2003
28 - 5
Technical issues of nearfield measurements inside a car for improving acoustic comfort
(a) Front lateral screen (10th octave band analysis)
(b) Rear lateral screen (10th octave band analysis)
Figure 4 : particular velocity on the lateral screen of the car
3.2. Sound Prediction inside the car The knowledge of the acoustic pressure and velocity fields on each side of the vehicle theoretically allows, thanks to the Helmholtz Integral Equation, to predict the pressure field at any point inside. Nevertheless, one difficulty is to model the exact Green function of the problem. In the considered case, some care have been taken to minimize the contribution of other side than the one of interest. This was done by setting up absorbing panels on each sides of the vehicle, except the side of interest. In that condition, the Green function is assimilated to the free field one. To scan the complete vehicle (2 lateral sides, windscreen and roof), no more than 1 hour was necessary : this definitely shows the operational relevance of the proposed system regarding the amount and richness of the produced outputs. The results of the prediction at two points inside the vehicle (at a center point and at the rear view mirror) are presented on figure 5 in comparison with the measurements at that points. Again, the predictions are very similar to the reference measurements. The observed differences in the high frequency domain at the rear view mirror point may be attributed to the simplicity of the propagation model (we assume a free field Green function). This expression can certainly be optimized by taking into account first order reflections for example.
Astelab 2003
28 - 6
Technical issues of nearfield measurements inside a car for improving acoustic comfort
(c) Sound pressure level inside the car at the center point (10th octave band analysis)
(d) Sound pressure level at the rear view mirror position (10th octave band analysis)
Figure 5: Prediction and measurement of the SPL inside the car
3.3. Numerical masking for component analysis Acoustical holography has a wide panel of treatment possibilities because it gives access to a complete description of the acoustic field (pressure and velocity). One major potentialities in the frame of car design domain is the prediction of the contribution inside vehicle of different components of interest. Consequently, it provides an operational tool which could be integrated to the entire vehicle conception process. On a practical point of view, this can be achieved by defining a numerical masking on the car body in order to identify the different components of interest. An illustrative example of such a functionality is presented on figure 6 which shows a complete numerical splitting of a car lateral side. In that example windows, components of the sealing system, doors and lateral mirror have been isolated and the associated acoustic transmission loss of each substructure has been calculated. The figure represents the difference of transmission losses between two sealing system configurations on the same car. The sealing system was changed between the two series of measurement. A 0 dB level on the figure indicates that no gain was obtained on the rear window. On the contrary, a deep change is to be noticed on the front one which gives an idea of the compared performances (in operational conditions) of the considered sealing system components.
Figure 6: Difference of transmission losses between two sealing system configurations on the same car. Blue colors indicate a gain in terms of acoustic isolation
Astelab 2003
28 - 7
Technical issues of nearfield measurements inside a car for improving acoustic comfort
4. CONCLUSION We presented the extended capabilities of nearfield acoustical holography in the domain of automotive industry, with a special attention to the localization and quantification in an operational context inside the vehicle. This work shows the relevance of the proposed approach in an industrial context in comparison with standard measurements methods. Especially the efficiency of the set-up and operation of the proposed approach has been demonstrated (less than 2 hours to scan the entire vehicle). Furthermore the LORHA system from METRAVIB RDS is offering dedicated post-processing tools that can produce a variety types of results of higher interest for the car industry. First, it allows to get the precise localization of acoustic sources on the ‘skin’ of the vehicle (pointing out examples as the window modal behavior as well as attenuation level coming from the sealing system). Then, it permits to calculate the partial acoustic contributions coming from the components to different positions inside the vehicle. Finally, it provides a useful tool for component design in the car conception phase. Sharing these kind of tools all along the engineering process, suppliers and OEMs can compare, with the same type of data, specifications of body components (transmission loss) from design to prototype validation, and even for laboratory measurements. Thus, it reduces numbers of component prototypes, and saves time and money as the dialog between suppliers and OEMs is getting clearer. In the future, the use of numerical methods (ray-tracing) from holographic raw data will allow to hear the influence of any virtual change of component as if the component itself was changed on the car.
REFERENCES 1.
J. Maynard, E.G. Williams, Y. Lee, Nearfield Acoustic Holography, Theory of generalised holography and development of NAH, J. Acoust. Soc. Am. Vol 78, n°4
2.
W.A. Veronesi, J.D. Maynard , Nearfield Acoustic Holography (NAH). II. Holographic Reconstruction Algorithms and Computer Implementation, J. Acoust. Soc. Am., 81(5), pp 1307-1322, may 1987
3.
B. Garnier, F. Molliex, D. Vaucher, The acoustic holography : from nice pictures… to effective applications, ICA95 – 15th International Congress on Acoustics, Acoustical Society of Norway – IUPAP, Trondheim, Norway, 26-30 June 1995, Vol.4, pp. 5560, M. NEWMAN Ed.
4.
D. Vaucher, D. Webster, B. Garnier, F. Molliex, MALICE, the efficient acoustical imaging system for precise noise localisation, 5th International Congress on Sound and Vibration, Adelaîde, Australia – Dec. 1997
5.
D. Vaucher, D. Webster, B. Garnier, F. Molliex, A new tool for sound proofing inspection : the SALSA system, 5th International Congress on Sound and Vibration, Adelaîde, Australia – Dec. 1997
6.
B. Garnier, D. Vaucher, J. Catalifaud, D. Bondoux, A new help for optimizing sound proofing packages : the SALSA system, 4th AIAA/CEAS Aeroacoustics Conference/19th AIAA Aeroacoustics Conference – AIAA/CEAS – TOULOUSE –2-4 juin 1998 – Paper 98 – 2244
7.
D. Vaucher, P. Mulocher, J. Catalifaud, JP. Demars, B. Florentz, F. Perrin , Holographie acoustique appliquée à des mesures intérieures d’habitacle 3D, SIA Conference – LE MANS – November 2000
8.
D. Vaucher, P. Chevret, F. Perrin , Use of Acoustical Holography in 3D Interiors measurements, Internoise 2002, The 2002 International Congress and Exposition on Noise Control Engineering, Dearborn, MI, USA. August 19-21, 2002
Astelab 2003
28 - 8
