Downloaded from SAE International by Univ of California Berkeley, Saturday, August 04, 2018

Evaluation of Vehicle Systems Structural Durability 2012-01-0953

Published
Using PSD Based Fatigue Life Approach 04/16/2012

Abhijit Londhe, Suhas Kangde and Narayana Moorthy

Mahindra & Mahindra Ltd.

Copyright © 2012 SAE International

doi:10.4271/2012-01-0953

the frequency domain with loads being specified in terms of a

ABSTRACT Power Spectral Density (PSD) function. This form of analysis
In current competitive environment automobile industry is is relatively quick. Transient analysis is performed in the time
under heavy pressure to reduce time to market. First time domain and requires significant computation time. This
right design is an important aspect to achieve the time and analysis is usually restricted to very short discrete events, and
cost targets. CAE is a tool which helps designer to come up is certainly not adopted when fatigue damage is attributable
with first time right design. This also calls for high degree of to long periods of random vibration loading. In that sense, the
confidence in CAE simulation results which can only be mission profiling approach helps the Virtual test. It allows
achieved by undertaking correlation exercises. you to simplify the applied load into single representative
PSD which is easier to handle, faster to process and shows a
In automobiles most of the structures are subjected to better statistical content than long and numerous time Series.
vibration from dynamic loads. All the dynamic road loads are
random in nature and can be very easily expressed in terms of The frequency domain PSD method has been used in space
power spectral density functions. In the current scenario and automotive industry for many years, for structures under
structural durability of the parts subjected to vibration is done simple loading conditions. In this paper a systematic PSD
partially through modal performance and partially though approach for automotive CAE durability analysis, dealing
frequency response analysis. The only question that arises is with multiple input loads and multiple road load events, is
what amplitude to use at what frequency and how to map all presented. The total damage and durability life of a structure
the accelerated tests dynamic load frequency spectrum to due to multiple random road load events, corresponding to
simulation domain. The paper describes a systematic and multiple road surfaces and events of a typical vehicle proving
economical approach of mapping all the accelerated tests in a ground test, are derived and implemented. A practical three-
single PSD and further using it for vibration fatigue. In step procedure for automotive CAE durability analysis using
addition a good correlation of simulation results with the test the random vibration approach is illustrated for system level
data is presented for the PSD based approach. structures under multiple road loading inputs. The procedure
incorporates (1) the proving ground (PG) road load data
process, (2) the finite element (FE) frequency response
INTRODUCTION simulation, and (3) the fatigue life evaluation.
In current competitive environment CAE is a tool for Comprehensive case study explaining the overall approach
verifying the durability of the design without making physical for an automotive bumper and its correlation with test is
prototypes. CAE durability can thus reduce time and cost for described in detail.
product development. The simplest analysis use a quasi-static
approach where the component is not sensitive to dynamic
effects such as resonance and transients. If dynamic effects
MISSION PROFILING
are important we must progress towards more intensive The useful life real world usage of the vehicle is mapped to
analysis such as random vibration, transient and modal the proving grounds by all automotive manufacturers.
transient analysis. Random vibration analysis is performed in Running the vehicle on the proving ground for small system
level changes is time consuming and costly. It was important
 Downloaded from SAE International by Univ of California Berkeley, Saturday, August 04, 2018

Figure 1. Calculating Fatigue Damage Spectrum

to find a method which can be used to validate the system obtained by summing the Fatigue Damage Spectra (FDS) for
level changes in test lab or a method to map proving ground all events over the time they are assumed to act. The FDS
test to simulation domain. With this as an objective mission calculated from a time signal estimates the likely damage
profiling was initiated. Mission profiling is a process of seen over that measurement period, therefore the FDS for
deriving a representative acceleration PSD that has atleast each time signal event should be scaled by the number of
same damage potential as the real environment. This times the event is likely in the lifetime of the vehicle. The
representative PSD then can be used by test lab for testing the total Lifetime fatigue damage is then found by summing the
system or by CAE department for analyzing the system. scaled FDS for each event as illustrated in Figure 2. a,b,c,d
are number of multiple runs on a given proving ground in
The ‘Fatigue Damage Spectrum (FDS)’ represents a plot of ADT cycle.
fatigue damage Vs natural frequency of a single degree of
freedom (SDOF) system. The Test Tailoring method assumes
that the dynamic characteristics of the final vehicle are
unknown. It takes a generic input acceleration and estimates
what acceleration values are likely to be witnessed by any
component mounted in that vehicle assuming a SDOF
transfer function between input and response. The natural
frequency of the system is also unknown so the method uses
a function to calculate the fatigue damage or shock for a
given range of natural frequencies. This ensures that all
eventualities are taken into account. Real world applications
are seldom SDOF systems, however, the response in most Figure 2. Mission FDS
cases is dominated by a single dynamic mode and this,
together with the inherent conservatism of the approach, is
Finally the mission FDS is inverted using same parameters
sufficient to ensure the synthesized test is at least as severe as
and mission PSDs are calculated. These mission PSDs are
that obtained in the real world. Figure 1 below shows the
representative of same damage potential as that of all proving
method used to calculate the Fatigue Damage Spectrum. The
ground tracks.
FDS is used to accumulate the damage caused by long term
exposure to fatigue damaging vibrations which, although
modest in amplitude, give rise to microscopic cracks that VIBRATION FATIGUE
steadily propagate over time and lead to eventual fatigue Dynamic road loads to automotive structures are random in
failure. nature and can therefore be conveniently defined as random
processes in terms of the PSD functions. PSD's are obtained
An accelerated test PSD is required that yields at least the by taking the modulus squared of FFT. In a PSD only the
same fatigue damage content as that seen by a component amplitude of sine wave is retained and all phase information
through its whole life. The total lifetime fatigue damage is is discarded. Typical multiple dynamic loadings are modeled
 Downloaded from SAE International by Univ of California Berkeley, Saturday, August 04, 2018

Figure 3. Mission Profiling for PSD

as a correlated PSD complex matrix. The random responses, The irregularity factor, α of PSD function, is defined as:
such as stresses and accelerations of an automotive structure,
are simulated efficiently by employing LMS Virtual lab
module. In frequency domain, fatigue damage of structures is
estimated based on the statistical properties of the response (5)
stress PSD function. Response stress PSD functions are
usually the results from finite element frequency response The material fatigue properties are usually measured as S-N
analysis. The most important aspect of vibration fatigue is to curve, which defines the relationship between the stress
understand the linking between the stress PSD and rainflow amplitude level SA versus the mean cycles to failure, N. For
cycle counting. Simplified form of derivation given below most high cycle fatigue problems the S-N curve can be
helps to understand this. defined as
The statistical characteristics of response stress can be
obtained through the moments of the PSD function. The nth
(6)
spectral moment of the stress PSD function S(f), frequency f
in units of Hz, is defined by the following equation: where B and m are material properties varying with loading
and environment conditions, such as mean stress, surface
finishing and temperature. The accumulated damage, E[AD],
due to fatigue random loading is evaluated based on the
(1) Palmgren-Miner's rule, and expressed as:

Properties of a continuous stationary Gaussian process can be

related to the above nth moments, mn of the PSD function.
The root mean square value, (7)

Where n(SA) is the number of cycles applied at stress

amplitude level SA, p(SA) is the probability density function
(2) of the stress amplitude. Substituting equation (6) into (7), a
general equation of fatigue damage from random stress
The average rate of zero crossing with positive slope, E[0], response is obtained.
which is also called as equivalent frequency in unit time, is
expressed as:

(8)
(3)
Where T is time duration of random loading. The function
p(SA) is usually defined as a function of statistical parameters
The average rate of peaks, E[p], in unit time is expressed as: of response stress PSD, such as E[0], E[p], α and mn. The
most popular damage estimating methods are Narrow band
method and Dirlik method. The total number of repeats, NT,

(4)
 Downloaded from SAE International by Univ of California Berkeley, Saturday, August 04, 2018

Figure 4. Accelerometers to measure base exciation in X,Y,Z direction on RH and LH side

Figure 5. Strain gauge locations on RH and LH side

of the fatigue life for the complete set of proving ground the expensive approach of running the full vehicle on
events is the inverse of the total damage. different proving ground for accelerated test validation. The
paper describes the methodology used to capture the failure
on the bumper system which is subjected to vibratory
loading. Typically bumper is mounted on the chassis and
subjected to dynamic loading during vehicle running
(9)
conditions.
The total life, TET, corresponding to the complete set of road
events, can be computed from the following equation. Bumper mounting bracket failure had occurred on the ADT.
Task was to capture the failure, suggest design modifications
and validate the modifications through simulation. Test
vehicle was instrumented to measure base acceleration in
X,Y,Z direction on the Chassis near the bumper mounting
bracket. Strain gauges were pasted near the failure location.
(10) Figure 4 & 5 below highlights the accelerometer and strain
In terms of accuracy and robustness, Dirlik empirical formula measurement locations on the vehicle. Instrumented vehicle
for rainflow ranges has been found to be far superior to the was driven on four different proving grounds for one pass and
previously available correction factors. data was acquired. This data was used for mission profiling
and generation of synthesized acceleration PSDs which are
representative of complete ADT cycle.
METHODOLOGY FOR EVALUATING
VEHICLE SYSTEMS FE MODEL, LOADS AND BOUNDARY
Mission profiling process helps in deriving the representative CONDITIONS
PSDs that has atleast same damage potential as the real
environment. Vibration fatigue process helps to calculate An automobile bumper is the front-most or rear-most part,
fatigue damage of structures based on the input PSDs from ostensibly designed to allow the car to sustain an impact
different tests. Important thing is to integrate the process to without damage to the vehicle's safety systems. They are not
meet the objective. Any vehicle system receives maximum capable of reducing injury to vehicle occupants in high-speed
loads when excited at its natural frequency. Vibration fatigue impacts, but are increasingly being designed to mitigate
helps to calculate the fatigue life for the such vibrating injury to pedestrians struck by cars. Typically bumper will be
systems. Vibration fatigue helps to validate system level mounted on the chassis and it will be subjected to dynamic
changes through simulation or using shaker table. This avoids loading during vehicle running conditions. When we design
 Downloaded from SAE International by Univ of California Berkeley, Saturday, August 04, 2018

such a system it is necessary to make sure that it should RESULT AND TEST CORRELATION
sustain all possible dynamic loads coming from road. Chassis
system transfers the road excitations to bumper. During test Bumper bracket failure image in ADT is shown in figure
all the excitations at the Chassis were captured using 3-axis 8(a). Vibration fatigue was able to capture the failure
accelerometer. Cut section of the Chassis was used in the FE location. In test the bracket failed in early ADT cycles.
model to excite the bumper system. Vibration fatigue also predicted the bracket failure in early
ADT cycles. Fatigue damage plot for the bumper bracket is
FE model consist of complete front bumper assembly and shown in the figure 8(b).
chassis cut section as shown in the figure 6. All bumper
bracket and chassis parts are modeled with shell elements.
Bolts are modeled by CBEAM elements. Seam welds are
simulated using rigid elements. Chassis cut section is
constrained to perform the modal analysis.

Figure 8(a). Bracket failure in ADT

Simulation stress PSD and test stress PSD were compared to

confirm the dominant frequency for the bumper. Figure 9
high lights the comparison between the test and CAE stress
PSD. Both top and bottom location dominant frequency
correlated with the test data. RMS comparison was done
Figure 6. Loads and boundary Conditions. between test and CAE at the top and bottom location of the
bumper mounting bracket. Good correlation of more than
90% was observed in the RMS values between test and CAE.
Modal analysis was performed on the bumper assembly to
understand the natural frequency of the system. First natural Design modifications were suggested based on the
frequency of the bumper came out to be around 32 Hz. deformations. These modification were evaluated using the
Second and third mode was bumper twisting and bumper same process. Modified design showed significant
bending modes. Synthesized PSD in X, Y and Z direction improvement compared to base design and cleared simulation
were used as input for the vibration fatigue. Figure 7 below acceptance criteria. Figure 10 below highlights the design
shows the overall process flow followed in CAE. modifications. Modified bracket cleared the physical ADT
cycles without any concern. Overall process has helped to
increase the confidence in the vibration fatigue approach.

Figure 7. Methodology for Vibration fatigue using mission PSD input

 Downloaded from SAE International by Univ of California Berkeley, Saturday, August 04, 2018

Figure 8(b). Damage per second plot for the bumper mounting bracket

values was observed. Modified design was evaluated using

the same process and cleared through CAE and test also.
Vibration fatigue analysis carried out with known loading is
more realistic than using the conventional 1g/3g FRF, since
the complete frequency spectra is mapped to model. This
method gives an idea of the life of the component and
therefore informs about the margin of safety. This approach is
conservative since it is based on stress life approach. Overall
exercise has helped Mahindra & Mahindra to increase the
confidence and improve the prediction capability for the
systems subjected to vibratory loading.

REFERENCES
1. Halfpenny, Andrew Dr.., Methods for accelerating
dynamic durability tests., 9th international conference on
recent advances in structural dynamics, Southampton, UK,
2006
Figure 9. CAE and Test Stress correlation
2. Bishop, Neil Dr., MSC Frimley and Alan Caserio, MSC
Costa Mesa., Vibration Fatigue Analysis in the Finite
Element Environment, Americas User Conference, Oct 5-9,
Sheraton Universal Hotel, Universal City, California
3. Halfpenny, A Dr.., Accelerated Vibration Testing Based
On Fatigue Damage Spectra, nCode International, 230
Woodbourn Road, Sheffield S9 3LQ, UK
4. Su, Hong Ph.D., Automotive CAE Durability Analysis
Using Random Vibration Approach, CAE Tools and Methods
Figure 10. Base and modified design
Group Advanced Technology Office, Visteon Corporation
5. nCode6.0 glyph works user manual
CONCLUSION 6. Halfpenny, Andrew., Environmental Accelerated Testing.
This paper has presented how the mission profiling approach Use of Virtual Test to Extend Physical Approach, nCode
can be clubbed with CAE based fatigue analysis to create an Whitepaper, Astelab 2008
accelerated and realistic test specification for a bumper
7. LMS Virtual Lab Online Help
system used in vehicles. Paper has also highlighted the
mission profiling methodology to arrive at the representative
random vibration profile. The obtained drive PSD was used
as input for a virtual FE-based fatigue analysis. A comparison
between fatigue lives obtained from test and CAE has been
performed and the results demonstrate good correlation.
These PSDs also helped to capture the failure on the bumper
bracket. Good correlation for the frequencies and RMS
 Downloaded from SAE International by Univ of California Berkeley, Saturday, August 04, 2018

CONTACT INFORMATION
Mr. Suhas Kangde
Mahindra & Mahindra (Automotive sector)
Plot No. 41/1, Anjur PO
Chengalpattu, Kanchipuram Dist.
Tamilnadu - 603204
Ph. No.: +91 44 22168101
Mob.:+91 9500017978

DEFINITIONS/ABBREVIATIONS
CAE
Computer Aided Engineering

RBE 2
Rigid bar element type 2

ADT
Accelerated Durability Testing

PSD
Power Spectral Density

SDOF
Single Degree of Freedom

The Engineering Meetings Board has approved this paper for publication. It has Positions and opinions advanced in this paper are those of the author(s) and not
successfully completed SAE's peer review process under the supervision of the session necessarily those of SAE. The author is solely responsible for the content of the paper.
organizer. This process requires a minimum of three (3) reviews by industry experts. SAE Customer Service:
Tel: 877-606-7323 (inside USA and Canada)
All rights reserved. No part of this publication may be reproduced, stored in a
Tel: 724-776-4970 (outside USA)
retrieval system, or transmitted, in any form or by any means, electronic, mechanical, Fax: 724-776-0790
photocopying, recording, or otherwise, without the prior written permission of SAE. Email: [email protected]
ISSN 0148-7191 SAE Web Address: http://www.sae.org
Printed in USA