|
Publications
| |
Digital Prototypes for Durability
From thakkar.raj(at)towerautomotive.com Wed May 27 16:14:21 1998
From: "Raj Thakkar"
Subject: Digital Prototype Text File
Raj Thakkar
Project Plan for SAE FD&E Task Group
"Digital Prototype for Structural Durability"
Vision, Mission, and Strategy are restated here with minor modifications
based upon the input received from Charlie Sieck-Caterpillar and other
members of the task group.
Vision: Develop methodology for using computer simulation to accurately
predict structural fatigue life prior to the first prototype build.
Mission: Demonstrate that the fatigue life of an ATV frame and
suspension components (load bearing structural components) can
accurately and repeatedly be predicted using a whole vehicle simulation.
Strategy:
* Develop recommended process for system simulation by openly sharing
data, methods, procedures, and results amongst members.
* Demonstrate that system simulation can be used for fatigue life
prediction.
* Extend current state of the art of fatigue life prediction to complex
specimens with complex loading.
* Develop systematic process for acquiring load inputs and processing
the data for application to a system simulation.
Success will be judged based upon how well the predicted fatigue life
compares with the life obtained from testing. It is suggested that a
factor of two be used.
Successful simulation of a system involves many phases. These phases are
outlined below as seen by Charlie Sieck and Raj Thakkar.
System Definition:
Geometry
* Accurately define geometry of all load bearing structural components,
i.e., frame, suspension components such as upper and lower control arms
(RH and LH)
* Define system topology showing locations of operator seat, tires
(include size), wheels, engine/transmission (in terms of mount
locations), rear axle, front and rear shocks (include shock angle), fuel
tank, battery, etc.
Boundary Conditions
* With and without tires
* With and without operators
* Fuel tank-empty or full
* any other boundary conditions to be used with FE simulation
* with and without engine
Component Identifications
* upper and lower control arms-LH & RH
* front-LH & RH and rear shocks
* frame
* engine/transmission
* wheel spindles
* fuel tank
* operator seat
* rear axle arm
* handlebar/stem
* tires size
Component Connections
* upper and lower Ball Joints-LH & RH
* front suspension bushings
* engine/transmission to frame mounting bolts
* steering linkage and its components
* shock to frame and shock to control arm
* suspension spring to frame and spring to control arm
General Information
* suspension Spring stiffness and its characteristics curve
* front and rear shock damping and its characteristics curve
* engine/transmission mount stiffness and damping
* front suspension bushing stiffness and damping
Load Inputs
* wheel spindle displacements or loads(X,Y,Z) at each wheel for all
types of road surfaces
* wheel spindle loads(X,Y,Z) due to severe braking and acceleration
* engine/drive line torque
* lateral loads due to severe turning or hitting a hard rock
* operator induced loads: (a) handlebars, (b) seat, (c) foot pegs20
* load or strain measurements at key control points for correlation with
FEA and for creating drive file for rig testing
Dynamic Simulation
Dynamic simulation depends upon how mass, stiffness, damping and force
quantities are defined and quantified.
* mass-lumped or distributed, static or dynamic mass which is frequency
dependent
* stiffness-rigid body or flexible, static stiffness or dynamic
stiffness which is frequency dependent
* damping-viscous, coulomb, or other, proportional vs non-proportional,
constant vs frequency dependent
* Quantification of the above properties depends upon whether static or
frequency dependent quantities are to be used. Frequency dependent
identification will provide closer dynamic simulation but is rather
difficult and time consuming.
Modal Testing and Correlation
* Through modal testing identify all primary modes and their associated
frequencies, i.e., first torsion, first vertical bending, first lateral
bending, and first parallelograming. Check for the orthogonality, and
purity of modes. Correlate mode shape and associated modal frequency for
each mode with FEA.
* identify boundary conditions and what is included in each measurement;
i.e. with or without engine, etc., free-free or supported on air bags
with stiffness-N/m.
* identify excitation technique used, i.e., hammer, shaker. If shaker is
used then identify input type, i.e., sweep sine, random, burst random,
etc.
* list all equipment type-include make, model and serial numbers
* FEA group must identify solver by name and version, type of algorithm,
e.g. lanczos, etc., and any numerical controls and checks used to
establish stability and accuracy
Static Testing and Measurements
* For all load bearing members calculate bend and torsional stiffness by
measuring load vs deflection, and torque vs twist angle (both stiffness
may not be required for all members). FEA and test group together must
identify structural members and determine which stiffness is required to
define each member.
* For certain members load vs strain data may be necessary. FEA and test
group together must define this requirement.
* Weigh each load bearing member, locate its centroid and mass center,
and calculate geometric moment of inertia and mass moment of inertia
Operational Inputs and Conditions
* operational inputs such as forces, accelerations, displacements,
strains, torque, and speeds, etc., associated with (a) different
maneuvers, (b) varied terrain, and (c) different operators must be
measured. Critical maneuvers, terrain, and operating conditions
associated with them must be identified before such measurements can be
made.
* above measurements must also be associated with vehicle configuration
such as fuel tank (full or partial), engine size, tire size, vehicle
speed and gear
Material Properties Data
* identify material for each load bearing component
* obtain all necessary static and fatigue data either through the
published information such as SAE J1099 or other means
* identify source of material data
* It is possible that material properties for each load bearing
components may not be readily and easily available. In that case
Material Properties Division will be asked to conduct search or test.
Rig Testing
* define vehicle configuration
* define boundary conditions, e.g. how vehicle is supported in the test
fixture
* define test condition, e.g. static, dynamic
* define load input including magnitude, location, direction
* identify source used in defining load data and controls used for
creating excitation file
* for multi-axis input identify phasing between loads
* for multi-wheel input, identify phasing between wheels and also
between loads
* provide clear description of test setup including pictures
* list test equipment and configuration including make, model and serial
numbers
* identify control parameters to be used
* identify location of points and type of output to be measured at each
location for correlation with FEA
Correlation Criteria
* correlate FEA vs experimental modal frequencies and associated mode
shapes
* correlate FEA calculated static load-strain (or displacement) data
with experimental measured static load-strain (or displacement) data on
rig
* correlate FEA calculated dynamic load-strain (or displacement) data
with experimental measured dynamic load-strain (or displacement) data on
rig
* establish criteria for acceptable correlation
* identify components and sub-assemblies for correlation and type of
correlation, e.g. strain, modal, to be used with each component and
sub-assembly
General Information-FEA
* define process and conditions for successful simulation
* list solver by name and version
* identify algorithm or algorithms used for each type of solution
* list boundary conditions and degrees of freedom for each type of
solution
* identify checks and criteria used to establish numerical stability and
accuracy
* list CPU requirements for each type of analysis
* list disk storage requirements for each type of analysis
* define data storage formats20
* define method for processing results
* list pre/post processor or processors by name and version
Real Time Simulation
If model is successfully correlated using simple input/output then the
attempt should be made to correlate it using real time input. In this
phase, correlation criteria should be based upon:
* comparing system response with system failure modes
* comparing component response with component failure modes
Note: need to establish failure modes and criteria
Fatigue Life Prediction
Fatigue life prediction should be an integral part of the overall
program. It should be coordinated by the FEA and the test groups with
the fatigue life prediction group so that life predictions are performed
and correlated at each stage of analysis. When program is completed, the
life prediction group should be able to:
* identify criteria for acceptable correlation between life predictions
based on FEA vs test data
* list significance of each variable on fatigue life prediction, e.g.,
boundary conditions, load input, etc.
* list criteria for improving correlation in future
* list process for performing fatigue life prediction without building a
prototype
Communication
Quick, easy and free transformation of the information between groups is
very important for the success of this program. Until a media for such
transformation is established, we must communicate through E-Mail and
telephone. Also, it is recommended that a copy of all the information
generated be sent to Raj Thakkar and Tom Cordes. I have started a file
which contains all the information sent to me on this program.
|
