Validation of Vehicle Nvh Performance Using Experimental Modal Testing and in-Vehical Dynamic Measurements
In:
Submitted By lunnm Words 4006 Pages 17
Licensed to Coventry University Licensed from the SAE Digital Library Copyright 2010 SAE International E-mailing, copying and internet posting are prohibited Downloaded Friday, October 22, 2010 6:54:13 AM
SAE TECHNICAL PAPER SERIES
2007-01-2320
Validation of Vehicle NVH Performance using Experimental Modal Testing and In-Vehicle Dynamic Measurements
Jennifer M. Headley, Kuang-Jen J. Liu and Robert M. Shaver
DaimlerChrysler Corporation
Noise and Vibration Conference and Exhibition St. Charles, Illinois May 15-17, 2007
400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-0790 Web: www.sae.org
Author:Gilligan-SID:4494-GUID:38672946-193.61.107.151
Licensed to Coventry University Licensed from the SAE Digital Library Copyright 2010 SAE International E-mailing, copying and internet posting are prohibited Downloaded Friday, October 22, 2010 6:54:13 AM
The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAE's peer review process under the supervision of the session organizer. This process requires a minimum of three (3) reviews by industry experts. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. For permission and licensing requests contact: SAE Permissions 400 Commonwealth Drive Warrendale, PA 15096-0001-USA Email: permissions@sae.org Tel: 724-772-4028 Fax: 724-776-3036
Licensed to Coventry University Licensed from the SAE Digital Library Copyright 2010 SAE International E-mailing, copying and internet posting are prohibited Downloaded Friday, October 22, 2010 6:54:13 AM
2007-01-2320
Validation of Vehicle NVH Performance using Experimental Modal Testing and In-Vehicle Dynamic Measurements
Jennifer M. Headley, Kuang-Jen J. Liu and Robert M. Shaver
DaimlerChrysler Corporation
ABSTRACT
NVH targets for future vehicles are often defined by utilizing a competitive benchmarking vehicle in conjunction with an existing production and/or reference vehicle. Mode management of full vehicle modes is one of the most effective and significant NVH strategies to achieve such targets. NVH dynamic characteristics of a full vehicle can be assessed and quantified through experimental modal testing for determination of global body mode resonance frequency, damping property, and mode shape. Major body modes identified from full vehicle modal testing are primarily dominated by the vehicle’s body-in-white structure. Therefore, an estimate of BIW modes from full vehicle modes becomes essential, when only full vehicle modes from experimental modal testing exist. Establishing BIW targets for future vehicles confines the fundamental NVH behavior of the full vehicle. In addition to vehicle body structure, the tire/wheel assembly, suspension, and chassis system affect overall on-road NVH performance of the vehicle. Understanding body acoustic and tactile sensitivity from various suspension and chassis attachment locations, and utilizing standardized NVH load cases under vehicle operating conditions, also benefit full vehicle NVH development. This paper presents modal test results of various vehicle body structures at the fully trimmed and body-inwhite configurations. Global body mode correlation between the two configurations is established, and the scatter band of the resonance frequency ratio between the two identified. In-vehicle NVH standardized test results are also presented to demonstrate the dynamic interaction of the vehicle to body structure modal behavior.
NVH refinement. The overall NVH performance of a vehicle contributes to a customer’s perception of a vehicles quality. A very quite, smooth riding car is considered to be well built. Thus, upfront planning for a vehicle’s NVH targeted performance is imperative. Once the desired performance level of the vehicle has been established, a detailed benchmarking study must be performed to establish NVH targets. These targets include, but are not limited to, the following: x x x x Global and Component Level Modal Mapping Attachment Stiffness and Acoustic Cavity Sensitivity Sub-system Performance Targets Vehicle Level Performance Targets for: x Powertrain Noise and Vibration x Road Shake x Harshness Feel and Boom x Idle Noise and Vibration x Windnoise
It will be demonstrated that the vehicle level performance is closely linked to the modal behavior of both the global and component level structure. Additionally, guidelines will be established to ensure optimal NVH performance of the system during the development phase before hardware is available.
MODAL BENCHMARING
One of the starting points to understanding a vehicle dynamic performance for noise and vibration is to understand the dynamic global structure modes. When beginning to work on a new vehicle program it is essential to understand the modal map. MODAL MAPPING It has been established that managing modes via a modal map is a necessity to avoiding potential modal interactions of a vehicles global and local modes with
INTRODUCTION
At the beginning of a new car program the attributes of the car are established; Such as, features including powertrain and driveline configurations, and desired
Licensed to Coventry University Licensed from the SAE Digital Library Copyright 2010 SAE International E-mailing, copying and internet posting are prohibited Downloaded Friday, October 22, 2010 6:54:13 AM
forcing functions such as road inputs and powertrain inputs.(1, 2) The traditional frequency ranges for powertrain, suspension and global modes are shown in an example of a program mode management chart in Figure 1.
Figure 1: Example of a Program Modal Map Distinct frequencies for all the key modes are chosen early in a program, and depend on known program information such as idle speed and the main engine orders. BODY IN WHITE MODAL The full vehicle global modes of interest, vertical bending and torsion, tend to be dominated by the vehicles BIW structure. Thus, analytical work tends to be done on the BIW model because it is of smaller size, and allows for quicker turnaround time on design iterations. Therefore, it becomes important to be able to establish the BIW global modal targets for use in program development. However, it is common that the BIW modes for the target vehicle are not known due to lack of a test specimen. Hence, it is important to establish the method of predicting the BIW modes based on the full vehicle test results. Establishing the Stiffness Ratio, K To establish the ratio of frequency difference between a BIW and full vehicle mode, eleven vehicles were tested at both configurations. A standardized modal test procedure was used with four independent un-correlated excitation force inputs and accelerometer measurements on all major components. The BIW and full vehicle mode shapes were visually inspected to insure consistency. The stiffness ratio, K, was obtained by dividing the BIW by the full vehicle frequencies for each mode classification; see Table 1 for a summary of the results. The average reduction in stiffness was 1.5 for vertical and lateral bending and 1.4 for torsion. In general, the ratio trended from 1.5 – 2.0. It should be noted that the lateral bending ratio is based on a smaller sub-set of data because this mode tends to be hard to establish in the full vehicle configuration. Table 1: Calculation of Average Stiffness Ratio, K COMPONENT LEVEL MODAL One of the most important component level targets included on the modal map are for the seat and steering column, which serve as the main vibration interfaces of the customer to the vehicle. Therefore, it is important that these modes are not coupled with major excitation forces. Development work of these components is typically done at the component level, thus the need for bedplate modal testing becomes critical. Steering Column Frequency Target A study of nineteen steering column assemblies was performed in an effort to develop a guideline for the column bending modes at the bedplate level. In addition, it was desired to develop the key design enablers necessary to have good modal performance of a column system in-situ. The test specimens were mounted to a rigid bedplate using their mounting attachment holes and the clamping mechanism of the steering column system. A control steering wheel and air bag module were used to eliminate the effect of the cantilevered mass from the design parameters. The vertical and lateral bending modes were identified using an impact hammer to excite the wheel in each direction, while measuring the corresponding frequency response functions (FRF). An example of the test set-up can be found in Figure 3. The benchmark system modes were used to establish a “ball-park” lower limit frequency target of 50 Hz (Figure 2). This number can be used as a rough guideline for the development of generic column architecture. It is believed that if met, this guideline will yield column architecture with enough “bandwidth” to be used in multiple programs. The individual program target value should be obtained using the mode management
Licensed to Coventry University Licensed from the SAE Digital Library Copyright 2010 SAE International E-mailing, copying and internet posting are prohibited Downloaded Friday, October 22, 2010 6:54:13 AM
strategy previously described, to develop the in-vehicle target, which will be cascaded down to bedplate level.
80 70 60 Frequency (Hz) 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 10 Column Vertical Bending Lateral Bending 11 12 13 14 15 16 17 18 19
Table 2: Design Parameters of Benchmarked Systems. Based on the correlation studies, assessment criteria, and the design enablers’ the following parameters were identified as the top three necessary to develop column architecture with desirable NVH performance: x x x A: Cantilever length from the column upper attachment to the steering wheel B: Cantilever length from the column aft-most bearing to the steering wheel C: Bracket attachment separation
Figure 2: Bedplate Modal Performance of Multiple Steering Column Systems. Top Column Design Enablers Excluding the steering wheel and air bag, a column system has ten key design elements that contribute to its modal behavior. See Figure 3 for detailed descriptions of these enablers.
Figures 4-6 contain the linear regression lines and 2 correlation coefficients (R ) for these three parameters.
340
R = 0.2491
"Canti-lever" Length from Attach to End Design Parameter A (mm) 320
2
300
280
260
Y=256
240
220 20 25 30 35 40 45 50 55 60 65 70 Ve rtical Be nding Freque ncy (Hz)
Figure 4: Design Parameter A as a Function of Column Vertical Bending. Figure 3: Test Set-up of Steering Column Testing on Modal Bedplate and Identification of Major Design Enablers. For this design study, eight of the ten parameters were considered. Items A-G and J were measured and tabulated in Table 2.
120 110 "Canti-lever" Length from Bearing to End Design Parameter B (mm) 100 90 80 70 60 50 40 30 20 20 25 30 35 40 45 50 55 60 65 70 Ve rtical Bending Fre que ncy (Hz)
2
R = 0.3801
Y=56
Figure 5: Design Parameter B as a Function of Column Vertical Bending.
Author:Gilligan-SID:4494-GUID:38672946-193.61.107.151
Licensed to Coventry University Licensed from the SAE Digital Library Copyright 2010 SAE International E-mailing, copying and internet posting are prohibited Downloaded Friday, October 22, 2010 6:54:13 AM
350
R = 0.2881
300 Attach Bracket Separation Design Parameter C (mm)
2
VEHICLE LEVEL NVH STANDARD LOAD CASES
Y=210
250
200
150
100
50 20 25 30 35 40 45 50 55 60 65 70 Ve rtical Be nding Freque ncy (Hz)
The use of standard NVH load cases is extremely important while developing a new vehicle. Helping to ensure comparable data over the span of the process, which can be years from target vehicle identification to production launch. There are six main standard full vehicle load cases that are used at the Chrysler group; powertrain noise and vibration, rough road shake, harshness, idle, coarse roadnoise, and windnoise. A key vehicle attribute, which is closely related to the vehicles modal map, is idle vibration at the seat and steering wheel. IDLE NOISE AND VIBRATION At idle, the primary forcing function is the idle frequency, hence, the vibrations felt at the seat and steering wheel are forced vibrations at this frequency. If a vehicle has a poorly executed modal map and the global vertical bending mode lies near the engine idle input frequency, it is expected to perform poorly. As the ratio of idle frequency to the vertical bending frequency, M, approaches one, the velocity of seat and steering wheel vibration should increase. To exemplify this theory, and establish the ratio values necessary to obtain good idle performance, multiple vehicles were measured for seat and steering wheel vibration at idle. The load case used for the study was the drive condition with no electrical load. Velocity measurements were made at the inboard drivers seat track attachment point, and the 12 o’clock position on the steering wheel, and the vector sum of the x, y and z spatial directions was calculated. Additionally, the corresponding idle frequency was recorded. Seat Track Vibration The linear regression analysis for the seat measurements can be found in Figure 8. For the measurements with a ratio less than one, vertical bending higher than the idle frequency, R2 achieves 65% correlation. When the ratio is greater than one, the function is 52% correlated. In both cases, seat shake is proven well correlated to the ratio, M. If it is desired to meet a generic idle vibration target at the seat track of 0.1 mm/s, the analysis indicates that M should meet the following guidelines:
Figure 6: Design Parameter C as a Function of Column Vertical Bending. The correlation coefficient between the vertical bending frequency and the top three design enablers was between 25 – 40%. The regression curves indicate to have a column that performs at the 50Hz level with minimum Cantilever length from the foremost attachment point to the end of the column, 256 mm should be the design criteria. Similarly, the system should have a Cantilever length from the bearing to the end of the column less than 56 mm, and a bracket attachment separation of at least 210 mm. It is acknowledged that correlation of the three design enablers individually is not sufficient to ensure the desired performance level. However, the correlation coefficient between the vertical bending frequency and certain combinations of the top three design enablers can be further improved. For example, the Cantilever length from the upper attachment to the shaft end plus two times of the Cantilever length from the bearing to the end has 55% correlation (Figure 7). Thus, each design enabler has different levels of influence to the vertical bending frequency of the steering column. The vertical bending frequency, as an objective function, can be further correlated to an optimal combination of all the design enablers, of which an even higher correlation can be achieved.
650
R = 0.5534
600
2
Design Parameter A+2*B (mm)
550
500
450
400
Y=368
350
300 20 25 30 35 40 45 50 55 60 65 70 Ve rtical Be nding Fre que ncy (Hz)
M < 0.8 or M > 1.76
(1)
Figure 7: Correlation of Combined Design Enablers A and B.
Author:Gilligan-SID:4494-GUID:38672946-193.61.107.151
Licensed to Coventry University Licensed from the SAE Digital Library Copyright 2010 SAE International E-mailing, copying and internet posting are prohibited Downloaded Friday, October 22, 2010 6:54:13 AM
Seat Shake at Idle R = 0.6599
0.6 Seat Vibration (mm/sec) 0.5
2
0.7
equations (1) and (2), it can be determined that the ratio of the idle to the vertical bending frequencies must be below 0.75.
CONCLUSION
The use of full vehicle and BIW level experimental modal testing determined the stiffness ratio between the two levels of hardware to be 1.5 for vertical and lateral bending, and 1.4 for torsion. This means that one can expect to see the modes drop by this ratio when comparing BIW to full vehicle. This rule of thumb can be very helpful in the design phase of a new vehicle. If the BIW test specimen is not available, the full vehicle modes can be used to calculate the BIW level targets. Allowing design iterations to take place at a simpler level, saving valuable design time. The optimal steering column subsystem generic modal target was demonstrated to be 50 Hz. Furthermore, the three main design enablers were identified as: the Cantilever length from the column upper attachment to the steering wheel, the Cantilever length from the column aft-most bearing to the steering wheel, and the bracket attachment separation. Individually, these three items were shown to correlate to the steering column bending frequency with R2 values of 20-40%. While the correlation of each characteristic was low, it was shown that combinations of these enablers could lead to further improved correlation. Vehicle level idle vibration performance was very clearly linked to the modal alignment of global vertical bending, idle frequency, and steering column resonance frequency. It was proven that to obtain desired seat shake performance, the ratio of the idle and vertical bending frequencies should be less than 0.8 or greater than 1.76. For desirable steering wheel vibration it was shown that the ratio should be less than 0.75 or between 1.24 and 1.38. Furthermore, to obtain a system with 1.0 and 0.6 mm/s idle vibration at the seat and steering wheel, respectively, the ratio should be less than 0.75. Future work is recommended to provide similar correlation between modal and other standard NVH load cases. One such study that would be very beneficial would be to correlate rough and smooth road shake a vehicles suspension or global torsion modes.
Ratio
Idle Frequency M ( Hz ) Vertical Bending Frequency
Figure 8: Seat Track Vibration Correlation to the Ratio of Idle and Vertical Bending Frequencies. Steering Wheel Vibration The linear regression curves for the steering wheel vibration as a function of the ratio M can be found in Figure 9. For ratios less than one, the correlation obtained was 57%, well correlated. When the ratio is above one a linear regression analysis will not sufficiently describe the behavior of the steering wheel vibration. This is because the behavior is a function of two variables, the vertical bending frequency and the first resonance frequency of the column. In general, the ratio of the columns first mode to bending is greater than one. Thus, as M becomes larger it approaches the first column mode. To account for this behavior a polynomial curve fit was used for this region. This allowed the correlation coefficient to reach 72%. The analysis reveals that for steering wheel vibration to achieve a generic target of 0.6 mm/s, the system should be designed for idle to bending ratios as follows: M < 0.75 or 1.24 < M < 1.38
3.0 2.5 Steering Wheel Vib ra tion (mm/sec) 2.0
Figure 9: Steering Wheel Vibration Correlation to the Ratio of Idle and Vertical Bending Frequencies. Optimized Ratio Value for Seat and Wheel Performance To develop a system that achieves a set idle vibration characteristic for the seat and steering wheel, the ratio values used for system targets must be must be optimized for both. Optimizing the ratio values found in
REFERENCES
1. Banner Tony, Deutschel Brian, Hamilton Dave, Juras Paul, “Development of the 2001 Pontiac Aztek Body Structure,” SAE 2000-01-1343, 2000. 2. Kim Ki-chang, Choi In-ho, Kim Chan-Mook, “A Study on the Development of a Body with High Stiffness,” SAE 2005-01-2464, 2005.
Licensed to Coventry University Licensed from the SAE Digital Library Copyright 2010 SAE International E-mailing, copying and internet posting are prohibited Downloaded Friday, October 22, 2010 6:54:13 AM
3. Dong B., Goetchius G., Duncan A., Balasubramanian M., Gogate S., “Process to Achieve NVH Goals: Subsystem Targets via “Digital Prototype” Simulations,” SAE 1999-01-1692, 2005.
BIW: Body-in-white, defined as the vehicle sheet metal processed through paint with fixed glass and all nonisolated structural bolt-on pieces installed. K: Stiffness ratio, defined as shown below for frequencies of a given mode shape.
CONTACTS
Jennifer Headley can be reached via email at JMH36@DCX.com Kuang-Jen Liu can KJL17@DCX.com Robert Shaver can RMS14@DCX.com be reached via email at
K
BIW FullVehicle
M: Ratio of idle frequency to the global vertical bending frequency. M = Idle Frequency/Vertical Bending Frequency R : Correlation coefficient
2
