Development of A Dynamic Vibration Absorber To Reduce Frame Beaming
Development of A Dynamic Vibration Absorber To Reduce Frame Beaming
Development of A Dynamic Vibration Absorber To Reduce Frame Beaming
Published 09/30/2014
Copyright © 2014 SAE International
doi:10.4271/2014-01-2315
saecomveh.saejournals.org
ABSTRACT
This paper describes the development and testing of a Dynamic Vibration Absorber to reduce frame beaming vibration in a
highway tractor.
Frame beaming occurs when the first vertical bending mode of the frame is excited by road or wheel-end inputs. It is
primarily a problem for driver comfort. Up until now, few options were available to resolve this problem.
The paper will review the phenomenon, design factors affecting a vehicle's sensitivity to frame beaming, and the principles
of Dynamic Vibration Absorbers (AKA Tuned Mass Dampers). Finally, the paper will describe simulation and testing that
led to the development of an effective vibration absorber as a field fix.
CITATION: Anderson, J., "Development of a Dynamic Vibration Absorber to Reduce Frame Beaming," SAE Int. J. Commer.
Veh. 7(2):2014, doi:10.4271/2014-01-2315.
381
382 Anderson / SAE Int. J. Commer. Veh. / Volume 7, Issue 2 (October 2014)
arise on smooth pavement when the natural frequency of this cab/sleeper, engine/transmission, cooling module and axles
mode coincides with wheel-end rotation frequency. Figure 1 were included as rigid bodies with the appropriate mass
shows an ODS (Operating Deflection Shape) of the frame properties. Tires are a simple linear model used for ride
beaming of a real bobtail tractor (without semi-trailer) at its comfort evaluation. More information about Kenworth's
problem speed. approach to full vehicle vibration simulation can be found in
reference [3].
Several things become clear from looking at the ODS: Figure 2. Finite element model of the test tractor. The hood and trailer
are not shown for clarity.
1. There are two nodal points for the mode (where there is
Since the model was constructed almost entirely of previously
no translational motion)
modeled components, only vehicle level validation was
2. Front cab mounts would ideally located near the forward considered necessary. Data were collected on the road
nodal point to minimize input to the cab. simulator for comparison to the vehicle model and to
3. The rear of the sleeper is far from the rear nodal point. A demonstrate that the frame beaming response was reasonably
suspension will be required for isolation. well reproduced. Model validation consisted of comparing test
4. Rear drive axle is far from a nodal point therefore it and simulation ODS and transmissibility functions from
will be very effective at exciting this mode. Wheel-end individual axle inputs to response points on the chassis. An
irregularities here must be minimized. example transmissibility function between rear drive axle
5. Drive axles move with the tail of frame, so suspension excitation and vertical response at B-pillar is shown in figure 3.
damping has little effect on this mode. Note that “background” 2 Hz sinusoidal symmetric excitation
was present at all other axle ends to ensure that the
Since the frame beaming mode is symmetric (left and right suspensions do not stick. These cause mismatch in
frame rails deform in phase), the tire irregularities that excite it transmissibility at low frequency between the test and
must also be symmetric. Frame beaming can come and go as simulation results.
the vehicle travels through curves. This is because the phase
angle changes between inside and outside wheels. Run to run The location of measurement points and some additional
results can vary. For this reason, much of the experimental work example measurement comparisons are given in the
described in this paper was performed on a road simulator, to appendix. Details of the validation work are outside of the
allow consistent phasing between left and right wheels. scope of this paper.
was not achieved with the sample mounts available. To look for
the “sweet spot,” simulation was run with the front engine mount
stiffness and damping varied over a wide range of values. Figure
8 below shows the result: The engine apparently does not work
as a vibration absorber. Rather than have the “sweet spot” that
was sought, there appears to be a “sour” one at about 1200 N/
mm and 0.5 Ns/mm. It is hypothesized that this is so because
the mass of the engine is much greater than that of the frame:
the frame becomes the absorber and has the higher amplitude.
As neither the test nor simulation yielded a useful solution, this
was not investigated further.
THE VIBRATION ABSORBER CONCEPT theoretically only limited by system damping. This is analogous
to our ladder frame's lightly damped first bending mode
The parameters studied within the current design do not offer a
(“beaming”) being excited by wheel-end input.
lot of improvement in frame beaming vibration. The
fundamental problem is still there: A lightly damped vibration
mode whose natural frequency can coincide with wheel-end
excitation in the normal operating range. There is no practical
way to add damping to the system as it stands nor to move the
natural frequency out of the operating range. This leads quite
naturally to the idea of changing the system dynamics with the
addition of a vibration absorber.
Theoretical Background
The concept of a vibration absorber is quite simple, and is
well described in reference [5]. A practical approach for
designing a vibration absorber is described in reference [6].
Here, only a quick overview is given as it pertains to the case
of a highway tractor.
The vibration absorber concept is best understood with simple Figure 8. Transmissibility vs. Frequency Ratio for Several Damping
lumped parameter models. Imagine a lightly damped, 1 degree Ratios
of freedom (DOF) system as shown if figure 7.
Suppose that one then adds a second, smaller mass/spring/
damper system to the original system as shown in figure 9.
Damping of the original system is assumed low and therefore
neglected. This is, after all, why we are adding a vibration
absorber in the first place. Damping of the beaming mode of a
typical tractor frame is low as evidenced by the sharp
resonance peak typically seen at the frame beaming frequency.
(1)
Where
Figure 9. System with Added Vibration Absorber (neglecting damping
is the frequency ratio (excitation/resonance) in the base system)
Where In the plot above, a damping ratio ζ2 of 0.2 gives a good result.
The trade-off is that cancellation is not as good at the original
is the resonant frequency ratio resonance frequency. So tuning of damping is a matter of
compromise. When properly tuned, one will replace the original
large resonance peak with two small, highly-damped ones.
is the mass ratio
Of practical concern is the sensitivity of this solution to tuning.
is the damping ratio for absorber alone Figure 11 shows the effects of the resonance frequency ratio
variation. As the system is detuned, the effectiveness of the
Recommended values for the mass and resonance frequency vibration absorber is much less. Notice that as the frequency
ratios are 0.2 and 1/1.2, respectively [5]. We will use these ratio is reduced, the higher frequency peak in the
values for the theoretical model. transmissibility function becomes more prominent. As it is
increased, the lower frequency peak becomes more prominent.
Without damping in the system, equation (2) simplifies to: This insight proved helpful when fine tuning the physical
vibration absorber on the real tractor on the road simulator.
(3)
Looking at the transmissibility, the expected two peaks are Figure 11. Transmissibility vs. frequency ratio with Added Vibration
Absorber for several resonance frequency ratios. ζ = 0.2, μ = 0.2
clearly seen when damping is low (figure 10). Between these
two peaks, very low transmissibility is possible with low
Figure 12 shows the effects of mass ratio on transmissibility. It
damping. This would be fine if the tractor only operated at this
appears that 0.20 is indeed the best ratio for this simple system.
frequency. Since it must operate across a range of frequencies
With a smaller mass ratio, the higher frequency peak becomes
(wheel rotation speeds), some damping is necessary.
more pronounced. With a larger mass ratio, the lower peak
becomes slightly higher but the upper peak is significantly
reduced. This comes at the penalty of increased weigh however.
Application to the Tractor The resulting B-pillar vertical acceleration with optimal tuning is
For a vibration absorber to be effective, it must be attached shown below (figure 14). Simulation predicts about a 75%
close to an anti-node for the mode shape (point of maximum reduction in peak RMS vibration with the vibration absorber
deflection). Put another way, if it were attached at a nodal point optimally tuned. From this, it was concluded that a physical test
for the mode shape, it will not affect the mode at all as the was warranted.
attachment point is not moving in that mode. Looking back to
the operating deflection shape of figure 1, there are three
candidate locations:
Figure 13. Contour plot of RMS acceleration at the driver's side B-pillar
for vibration absorber stiffness and damping with rear drive tire run-out
for speeds 80 to 100 km/h with worst-case tire phasing, without
semi-trailer
388 Anderson / SAE Int. J. Commer. Veh. / Volume 7, Issue 2 (October 2014)
SUMMARY/CONCLUSIONS
This paper reviewed parameters of the current design thought
to affect frame beaming vibration in the cab, and the
development of a dynamic vibration absorber to alleviate that
vibration.
CONTACT INFORMATION
The author may be contacted at:
Johnbrian.anderson@yahoo.com
(Home)
ACKNOWLEDGMENTS
I would like to thank John Olsen, Donald Smith, and the staff of
the PACCAR Technical Center for their support in this
investigation. I would also like to thank Kenworth R&D for
excellent design support, fabrication and installation of the
proof of concept; and Simon de Cock, René Liebregts, and my
colleagues at DAF Trucks in the Netherlands for their input and
advice during this investigation.
390 Anderson / SAE Int. J. Commer. Veh. / Volume 7, Issue 2 (October 2014)
APPENDIX
Measurement points are numbered below. Triangles indicate assumed rigid bodies (cab/sleeper, engine, cooling module, tanks, etc.).
These were used for comparing transmissibility functions and operating deflection shapes on the road simulator.
Anderson / SAE Int. J. Commer. Veh. / Volume 7, Issue 2 (October 2014) 391
Below is a comparison of transmissibility functions between rear drive axle symmetric excitation (at tire contact points) and responses
on vehicle in vertical (3) direction for the tractor without trailer. Blue is from test, red is from simulation. Note that “background” 2 Hz
sinusoidal symmetric excitation was present at all axle ends to ensure that the suspensions do not stick. These cause the excessive
peak in the test data. These plots are given as an indication of model quality. Responses to symmetric and antimetric excitation at all
axles with and without loaded semi-trailer were examined but are not given in this paper.
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 International.
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the
paper.