Operational Modal Analysis of tyre road interaction using

Abaqus Explicit and Operational Modal Analysis

P. Sakthivel1, K. V. Narasimha Rao2, R. Krishna Kumar1
1
Indian Institute of Technology Madras, Engineering Design Department
Chennai 600036, Tamilnadu, India
e-mail: [email protected]
2

JK Tyre and Industries Ltd., RPSCOE for Tyre & Vehicle Mechanics,
IIT Madras, Chennai 600 036, Tamilnadu, India.

Abstract
Modern vehicles have well established control mechanisms for power unit noise reduction. On the other
hand, noise due to tyre road interaction has become a significant contributor of vehicle interior and
exterior noise. Understanding tyre structural dynamics is very important to mitigate the tyre/road noise
problem. The objective of this work is to establish a procedure to acquire vibration response data for
Operational Modal Analysis (OMA), using a dynamic finite element (FE) analysis. This has been
achieved in two stages. Firstly, the experimental procedure is replicated using Abaqus Explicit, a
commercial dynamic FE analysis software. Secondly, the tyre dynamic characteristics are determined
through Operational Modal Analysis by integrating the explicit FE analysis results with LMS Test Lab
software. The modal parameters of the rolling tyre are extracted by the use of polyreference least squares
complex frequency domain (LSCF) approach.

Introduction

Tyre/road interaction is the main source for tyre vibrations. The spindle force due to this interaction
attains its maximum value at the occurrence of structural resonance of tyre wheel assembly. It is further
transmitted to the vehicle body through the suspension sub system. This body force excitation is
responsible for structure borne interior noise (for a comprehensive account of tyre noise, refer to Sandberg
and Ejsmont [1]).
Modal frequencies, damping and mode shapes describe the dynamic behaviour of a tyre. Traditionally
these quantities are found from Experimental Modal Analysis and/or finite element analysis of a nonrolling tyre with the basic assumption of system linearity and reciprocity. Extraction of these quantities is
obtained from the estimation of Frequency Response Functions for the known dynamic excitation force.
In reality, it is impractical to measure the excitation force of a rolling tyre from the tyre road interface.
Under this situation, Operational Modal Analysis (OMA) can be used for modal analysis [2].
The modal parameters obtained from the Experimental Modal Analysis of a stationary tyre may be
modified during rolling, though the static mode shapes are useful to understand the mode shapes obtained
from the Operational Modal Analysis. Excitation of a rolling tyre over a cleat gives relatively a flat
spectrum and is preferred for the measurement of acceleration responses.
The procedure developed in this paper is useful to carry out a parametric study to understand the influence
of driving speed, cleat dimension, inflation pressure, static preload etc.
Peter Kindt et al. [3,4] have extracted modal parameters of a rolling tyre, using tyre on tyre over a circular
cleat. Laser Doppler Vibrometer was used to measure the acceleration response of the smooth tyre at
various points on the tread and sidewall and used as an input for Operational Modal Analysis.

1603

1604

P ROCEEDINGS OF ISMA2012-USD2012

Peter Willem Anton Zegelaar [5] showed that the modes of vibration of tyre strongly depend on tyre
construction, rotational velocity, spindle boundary conditions and contact patch boundary
conditions.Modal parameters are extracted from the measured frequency response functions that describe
the dynamic response of the structure to a known dynamic excitation force. The tyre is excited at one
point and the acceleration responses are measured in several points on the tread and sidewall and wheel.
L. H. Yam et al. [6] have obtained three dimensional mode shapes (radial, tangential and lateral directional
modal parameters) of a tyre using Experimental Modal Analysis.
Operational Modal Analysis provides a complete modal model under operating conditions. Sum of
Crosspower spectrum is used to extract the modal parameters of rolling tyre. T. Lauwagie et al. [7] have
done a comparative study of modal parameter extraction procedure viz., Operational Modal Analysis,
Experimental Modal Analysis and combined Operational and Experimental Modal Analysis on a small
hydraulic crane modal analysis study.
Modal analysis results of a tyre have significant contribution in tyre/road interaction noise prediction.
Tyre modal data are used directly in Boundary Element Method and Infinite Finite Element Method based
tyre models. Nakajima et al. [8] have used the tyre modal analysis results to define the surface vibrational
boundary conditions for predicting air borne noise fields. Michel Constant et al. [9] have studied the
vibro-acoustic interaction between tyre and car sub systems and identified the main suspension parts
affecting the structure borne interior noise. Operational deflection shapes of tyre wheel subsystem were
used by Ichiro Kido [10] to quantify the force transmitted through suspension sub system to the vehicle
body, to address the structure borne interior noise of the vehicle. In-spite of the importance of the rolling
tyre modal analysis, very few attempts has been made in this direction.
There are various modal identification techniques for extracting the modal parameters using Operational
Modal Analysis [11,12]. They are grouped as follows: (1) Time Domain approach, (2) Frequency Domain
approach and (3) Time-Frequency plane approach. The time domain approach includes Natural Excitation
Technique (NExT), Auto Regressive Moving Average technique (ARMA) and Stochastic Subspace
Identification technique (SSI). The frequency domain approach includes Frequency Domain
Decomposition (FDD) and Enhanced Frequency Domain Decomposition (EFFD) techniques. Time
Frequency plane is the Wavelet Transform Technique. In the present work LMS Test Lab software
implementation of PolyMAX algorithm has been used to extract the modal parameters. Initially the
method has been developed for classical modal analysis that uses measured frequency response functions
as primary data and the theory behind the method can be found in P. Guillaume et al. [13] and B. Peeteers
et al. [14]. Later it has been proved by B. Peeteers et al. [15] that the LMS PolyMAX estimator can also
be used effectively in Operational Modal Analysis. The following section describes the finite element
modeling of tyre and the procedure of integrating the results of finite element analysis with LMS Test Lab
software to carry out the Operational Modal Analysis.

Finite element model and modal extraction

Rao K. V. N. et al. [16] have used a finite element tyre model to study cornering, braking and corneringcum-braking dynamic behaviour of tyre. The detailed model, based on a hyperelastic material, includes
all the reinforcements of the tyre. Their approach is adopted to develop a finite element model of a
205/55R16 tyre.
Explicit dynamic rolling of finite element model over a semicircular cleat of 25 mm radius has been
carried out. This numerical simulation replicate the experiments followed by Peter Kindt et al. [3,4]. The
same inflation pressure of 220 kPa and driving speed of 28.3 kmph are used to compare the results.
Figure 1a shows the axi-symmetric half finite element tyre model that has been used to generate the full
tyre model. The colour group describes the various components and constructional details of the tyre.
The tyre includes two belts and one ply and one protection ply.
Figure 1b shows the 3D finite element model, rolling over a semi-circular cleat. Smooth tyre has been
considered to eliminate the influence of tread pattern vibrations and tread groove resonances over the
whole tyre vibrations. The spindle force and tyre structural vibrations responses were obtained from the

FP7 IAPP TIRE-DYN T YRE / ROAD NOISE AND EXPERIMENTAL VALIDATION

1605

simulation results. These measurements were made with the sampling frequency of 2048 Hz. The
spectral content of the spindle force were obtained from its power spectral density function.

Figure 1: (a) Axi-symmetric half tyre model

Figure 1: (b) Tyre rolling over a cleat

The numerical simulation has been completed in two stages. In the First stage, the inflation and loading of
the finite element tyre and rim assembly has been simulated. The inflation pressure of 220 kPa was
defined and a pre load of 2930 N was given by lifting the road, which is modeled as a rigid surface. All
the six degrees of freedoms (dof) of the tyre rim spindle point were arrested. All degrees of freedom
except the vertical translation of the rigid road were arrested. In the second stage, rolling of the tyre over a
cleat has been simulated. The following boundary conditions were ensured. Except the longitudinal
translation and rotation about spindle axis of the tyre rim assembly, rest of the dofs were arrested. For the
road with the lifted position all the degrees of freedom were arrested to simulate the straight rolling of
tyre. The acceleration responses in x, y and z directions with respect to the global Cartesian reference, at
240 points in four circles of tyre, two on tread and one each on right and left side wall respectively were
written as an output during the simulation.
These response results were communicated to the LMS software by defining appropriate header
information that describes the data characteristics, sampling rate and number and name of the channels
and database information. Thus the data input to the software is given as if they have been acquired
through data acquisition system and stored as throughput data format. The tyre geometry was created
using the 240 measurement points as shown in Figure 3. Tyre right hand tread leading edge node point
named RTD72074 is highlighted in the Figure 3. The vertical direction acceleration response of this node
was chosen to be the reference signal for auto and crosspower function calculations.

Figure 3: Tyre Modal Model Geometry

1606

P ROCEEDINGS OF ISMA2012-USD2012

The auto and crosspower functions of output responses were determined for the modal parameter
extraction. Further the LMS Test Lab implementation of PolyMAX algorithm in frequency domain was
used to formulate the least squares problem to estimate the mode shapes. The advantage of using the
frequency domain implementation is that the start and end frequencies can be specified and that residual
effects of modes outside the band can be taken into account. Also the closely spaced poles are well
identified. Crosspower function plays a similar role in Operational Modal Analysis as that of Frequency
Response Function in case of Experimental Modal Analysis. Crosspower sum of all the signals is shown
in Figure 4. The peaks in the crosspower sum represent the modal frequencies and the sharpness of the
peaks decides the corresponding damping value.

Figure 4: Crosspower sum

Curve fit of the sum of all the crosspower functions for the required modal orders has been obtained from
the stabilization diagram. The selection of stable poles that gives the correct modal parameters from the
stabilization diagram is a challenging task. Selecting the poles by visual inspection requires experience
and skill, as the selection is based on similarity in frequency, damping ratio and/or mode vector between
poles belonging to subsequent modal orders. Hence LMS Test Lab Automatic Modal Parameter Selection
(AMPS) tool was used to select the stable poles.
Figure 5 shows the stable poles picked from the stabilization diagram for the frequency band of 2 Hz to
250 Hz and the modal order of 106. The vertical lines represent the physical stable pole. There are peaks
that are not picked by AMPS tool between 67 Hz to 84 Hz as they represent unstable mathematical modes.
The main advantage of using the AMPS tool is that it selects only the correct physical stable poles for any
number of repetition of modal picking on the stabilization curve and gives user independent results. Thus
the Operational Modal Analysis of tyre road interaction of finite element model of 205/55R16 tyre has
been completed using LMS Test Lab Operational Modal Analysis Module.

FP7 IAPP TIRE-DYN T YRE / ROAD NOISE AND EXPERIMENTAL VALIDATION

1607

Figure 5: Stabilization diagram and pole selection by AMPS

Figure 6a shows the contact pressure distribution just before the cleat impact. The symmetric contour
profile describes the steady state straight rolling of tyre and Figure 6b shows the contact pressure
distribution at the instant of cleat impact and it represents the impulse excitation which is responsible for
flat excitation spectrum.

Figure 6: (a) Contact pressure of rolling tyre

Figure 6: (b) Contact pressure during cleat impact

Figure 7 shows vertical spindle force excitation of the rolling tyre and it is clearly seen that the impulse
excitation peak resulted due to the cleat impact excitation.

1608

P ROCEEDINGS OF ISMA2012-USD2012

Figure 7: Vertical Spindle Force due to cleat impact excitation

Results and discussion

Figure 8 shows acceleration response in x, y & z directions of a node point that was located on the right
side leading edge tread of the tyre. The peak response of acceleration at 1.04sec clearly shows the instant
of the cleat impact on the tyre. Minimum response of lateral acceleration ensures the straight rolling of
tyre and shows the insignificant excitation in lateral direction. The periodic zero response in all three
directional responses represent the tyre tread contact point with the road during every revolution.
Figure 9a shows the spindle force response due to cleat impact on rolling tyre in all the three directions.
These plots are plotted from just before the cleat impact instant. The vertical spindle force includes both
static pre-load and the dynamic response force due to cleat impact. The longitudinal and lateral spindle
forces are the dynamic response forces alone. Dynamic spindle force response in lateral direction states
the negligible energy content of lateral excitation.
The frequency content of these spindle forces, shown in Figure 9b was obtained from their power spectral
density functions. The correlation of the spectral content of the spindle force with the modal frequencies
validates the results. The power spectral density plots of vertical and longitudinal spindle force represents
significant energy content up to 250 Hz, which is responsible for structure borne interior noise. The effect
of power spectral density function of lateral force to structure borne interior noise needs to be investigated.

FP7 IAPP TIRE-DYN T YRE / ROAD NOISE AND EXPERIMENTAL VALIDATION

Figure 8: Acceleration response of a node point in X, Y & Z directions

obtained from Finite Element Analysis

1609

1610

P ROCEEDINGS OF ISMA2012-USD2012

Figure 9: (a) Spindle force excitation

Figure 9: (b) PSD of spindle force excitation

Figure 10a and Figure 10b shows the correlation obtained between the original operational crosspower
spectra and the synthesized crosspower spectra of well correlated signal for two different inflation
pressure respectively. The error between the spectra did not affect the stable poles selected. For all the
extracted frequency values very good correlation between the spectra is obtained. Also it gives an
immediate visual assessment of the accuracy of the modal parameter estimation procedure.

FP7 IAPP TIRE-DYN T YRE / ROAD NOISE AND EXPERIMENTAL VALIDATION

1611

Figure 10: (a) Comparison of operational and synthesized crosspower for tyre inflation pressure 220kPa

Figure 10: (b) Comparison of operational and synthesized crosspower for tyre inflation pressure 270kPa

The results of the modal parameters and the operational mode shapes are arranged in Figure 11. Since the
measurement points of acceleration responses obtained from the finite element model are more than the
measurement points considered in [3], structural resonance frequency upto (10,0), the circumferential belt
modes are obtained. These results are extracted for a driving speed of 28.3 kmph which corresponds to
the tyre free rolling speed of 25 rad/sec, inflation pressure of 220 kPa and the static preload of 2930 N.

1612

P ROCEEDINGS OF ISMA2012-USD2012

Figure 11: Operational mode shapes of 205/55R16 tyre due to tyre/road interaction
The modal frequencies and damping values of the current results are compared with the published results
of Peter Kindt et al. [3] in Table 1. Though the mode shapes match, the frequency and damping values do
not match. The variation in the natural frequency values may be due to the difference in the geometry and
properties of the tyre. It is proposed that these limitations will be overtaken in the ongoing research work.

FP7 IAPP TIRE-DYN T YRE / ROAD NOISE AND EXPERIMENTAL VALIDATION

Mode Shape
Mode: (0,0)
Mode:(1,1)horiz

Published results
Frequency
Damping
(Hz)
(%)
*
*

1613

Current results
Frequency
Damping
(Hz)
(%)
4.067
6.34
*
*

77.15

6.01

Mode:(1,0)vert

81.83

4.34

59.933

0.38

Mode: (2,0)0

102.98

3.02

Mode: (2,0)extr

113.94

3.66

84.733
*

0.3
*

Mode: (2,1)0

96.09

3.41

Mode: (2,1)extr

91.29

2.79

Mode: (3,0)0

124.09

2.6

96.791

0.33

Mode: (3,0)extr

139.33

3.87

109.124

0.06

Mode: (4,0)0

151.76

2.62

116.956

0.28

Mode: (4,0)extr

167.52

3.29

129.534

0.14

Mode: (5,0)0

182.7

3.71

133.41

0.18

Mode: (5,0)extr

199.88
*

1.5
*

141.44

0.14

161.859

0.31

Mode: (6,0)extr

170.993

0.23

Mode: (7,0)0

186.817

0.18

Mode: (7,0)extr

195.055

0.21

Mode: (8,0)0

199.18

0.26

Mode: (8,0)

210.699

0.35

Mode: (8,0)extr

218.567

0.23

Mode: (9,0)0

224.237

0.32

Mode: (9,0)

228.929

0.33

Mode: (9,0)extr

232.632

0.29

Mode: (10,0)

241.992

0.25

Mode: (6,0)0

* indicates missing modes and published results are taken from Peter Kindt et al. paper [3]
Table 1: Results of modal parameters of 205/55R16 tyre rolling over 25mm semi circular cleat
In order to understand the influence of inflation pressure the same procedure was repeated for inflation
pressure of 270 kPa keeping the same driving speed, cleat dimension and static deflection. The radial
preload is increased due to the inflation pressure as the static deflection is kept a constant. There is a
positive frequency shift of the structural resonances observed due to increase in inflation pressure. The
shift is resulted due to increase in tyre dynamic stiffness. There is also a noticeable change in damping
values due to inflation pressure variation. The modal parameters values of these two cases are tabulated in
Table 2.
From Table 2 it is observed that there is a 6 to 15 Hz positive frequency shift of structural resonance
frequencies due to an increase in the inflation pressure. This is due to an increase in the belt and sidewall
stiffness due to higher inflation pressure. This trend matches with the results published by Peter Kindt et
al. [3,4]. Also it is observed that there is a rise in radial preload of the tyre. This is due to the fact that in
the finite element simulation the road lift is maintained for both inflation pressure cases to have the same
static deflection of the tyre.

1614

P ROCEEDINGS OF ISMA2012-USD2012

Mode Shape
Mode: (0,0)
Mode:(1,1)horiz

Inflation pressure 220 kPa

Frequency
Damping
(Hz)
(%)
4.067
6.34
*
*

Inflation pressure 270 kPa

Frequency
Damping
(Hz)
(%)
4.0593
5.87
*
*

Mode:(1,0)vert

59.933

0.38

65.5958

0.47

Mode: (2,0)0

84.733
*

0.3
*

90.6120
*

0.36
*

Mode: (2,1)0

Mode: (2,1)extr

Mode: (3,0)0

96.791

0.33

102.6674

0.23

Mode: (3,0)extr

109.124

0.06

Mode: (4,0)0

116.956

0.28

Mode: (4,0)extr

129.534

0.14

139.2683

0.12

Mode: (5,0)0

133.41

0.18

143.2273

0.07

Mode: (5,0)extr

141.44

0.14

151.3242

0.10

Mode: (6,0)0

161.859

0.31

173.7737

0.28

Mode: (6,0)extr

170.993

0.23

182.3514

0.27

Mode: (7,0)0

186.817

0.18

201.6263

0.25

Mode: (7,0)extr

195.055

0.21

209.0481

0.29

Mode: (8,0)0

199.18

0.26

214.5045

0.10

Mode: (8,0)

210.699

0.35

216.9282

0.19

Mode: (8,0)extr

218.567

0.23

225.2255

0.21

Mode: (9,0)0

224.237

0.32

232.6206

0.31

Mode: (9,0)

228.929

0.33

Mode: (9,0)extr

232.632

0.29

239.1948

0.07

Mode: (10,0)

241.992

0.25

Mode: (2,0)extr

* indicates missing modes

Table 2: Results of modal parameters for different inflation pressures

Conclusion

This paper demonstrated the procedure developed for conducting the Operational Modal Analysis of
tyre/road interaction using Abaqus explicit finite element code and LMS Operational Modal Analysis.
The repeatability and accuracy of the simulation procedure leads the way to do parametric study
effectively. In this paper only the influence of inflation pressure on tyre vibration characteristics has been
presented. It is proposed to carry out detailed parametric study of the influencing factors affecting the
operational modal parameters in the ongoing research work.

Acknowledgements
Research work was supported by Raghupati Singhania Centre of Excellence for Tyre & Vehicle
Mechanics, sponsored by JK Tyres and Industries at IIT Madras, Chennai, India.

FP7 IAPP TIRE-DYN T YRE / ROAD NOISE AND EXPERIMENTAL VALIDATION

1615

References
[1] U. Sandberg, J. A. Ejsmont, Tyre/road noise reference book, Informex, Kisa, Sweden,2002.
[2] Mehdi Batel, Operational Modal Analysis - Another way of doing modal testing, Sound and
Vibration, August 2002, pp 22-27.
[3] P. Kindt, F. De Coninck, P. Sas, W. Desmet, Analysis of tire/road noise caused by road impact
excitations, Noise and Vibration Conference and Exhibition St. Charles, Illinois May 15-17, 2007,
SAE paper 2007-01-2248.
[4] P. Kindt, D. Berckmans, F. De Coninck, P. Sas, and W. Desmet. Experimental Analysis of the
Structure-Borne Tyre/Road Noise due to Road Discontinuities, Mechanical Systems and Signal
Processing 23 (2009) 2557-2574.
[5] Peter Willem Anton Zegelaar, The dynamic response of tyres to brake torque variations and road
unevennesses, PhD Thesis, 1998, Delft University of Technology, The Netherlands.
[6] L. H. Yam, D. H. Guan and A. Q. Zhang, Three dimensional mode shapes of a tire using
experimental modal analysis, Experimental Mechanics, 2000, pp 369-375.
[7] T. Lauwagie, R. Van Assche,J. Van derStraeten, W. Heylen, A Comparison of Experimental,
Operational, and Combined Experimental-Operational Parameter Estimation Techniques,
Proceedings of the International Noise and Vibration Conference, ISMA 2006, Leuven, Belgium,
September 2006.
[8] Nakajima, Y., Inouc, Y., Ogawa, H., Application of the Boundary Element Method and modal
Analysis to Tire Acoustics Problems, Tire Science and Technology, TSTCA, Vol. 21 No. 2, AprilJune, 1993, pp 66-90, presented 1992 at The Tire Society Conference.
[9] M. Constant, J. Leyssens, F. Penne, R. Freymann, Tire and car contribution and interaction to low
frequency interior noise, SAE 2001 Noise and vibration Conference; Traverse City, MI., 2001 May,
SAE paper 2001-01-1528.
[10] Ichiro Kido and SagiriUeyama, Coupled Vibration Analysis of Tire and Wheel for Road Noise
Improvement, SAE paper 2005-01-2525.
[11] LMS Test Lab Operational Modal Analysis Theory Documentation, LMS International.
[12] Nis Moller, SvendGade, Resent Trends in Operational Modal Analysis, SAE paper 2003-01-3757
[13] P. Guillaume, P. Verboven, S. VanLanduit, H. Van derAuweraer and B. Peeteers, A poly-reference
implementation of the least squares complex frequency domain estimator, Proceedings of the IMAC
21, the International Modal Analysis Conference, Kissimmee (FL), USA, February 2003.
[14] B.Peeteers, P. Guillaume, H. Van derAuweraer, H. Cauberghe, P. Verboven, and J. Leuridan,
Automotive and aerospace applications of LMS PolyMAX modal parameter estimation method,
Proceedings of the IMAC 22, the International Modal Analysis Conference, Dearborn (MI), USA,
January 2004.
[15] B. Peeteers, and H. Van derAuweraer, PolyMAX: a revolution in Operational Modal Analysis,
Proceedings of IOMAC, the 1st International Operational Modal Analysis Conference, Copenhagen,
Denmark, 26-27 April 2005.
[16] Rao, K.V.N., Kumar,R.K., and Bohara, P.C., Transient Finite Element Analysis of Tire Dynamic
Behavior, Tire Science and Technology, TSTCA, Vol.31, No.2 April-June 2003, pp.104-127.

1616

P ROCEEDINGS OF ISMA2012-USD2012