Simulation vs. Testing: Getting the Best of Both Worlds T. Kim Parnell, Ph.D., P.E. Janet Neo PEC – Parnell Engineering & Consulting www.parnell-eng.com
Simulation vs. Testing: Getting the Best of Both Worlds
T. Kim Parnell, Ph.D., P.E.Janet Neo
PEC – Parnell Engineering & Consultingwww.parnell-eng.com
IntroductionSimulation vs. TestingWhat are the issues?
Benefits of Synergizing Simulation and TestingIllustrations & Case StudiesConclusionsQuestions??
Outline of Presentation
Introduction
• Finite element analysis (FEA) and physical testing are complementary
• A comprehensive program needs to include both components
• With judicious experimental validation, FEA can be used to reduce the amount of physical testing that is needed and shorten the design cycle
The Challenge for Medical Device Development
• Reduce development time• Increase confidence of success• Avoid surprises and delays
Prototype Development
1. Physical prototypeCost and lead time is often a limitationEssential for animal testing and determining needed characteristicsWant to reduce the number of design iterations that are prototyped
2. Virtual prototypeAssess more design options Compare alternatives
Testing Is Essential for:
• Detailed characterization of the material; getting data needed for the analysis
• Fatigue testing taking into account surface finish, processing steps
• Validation
Sensitivity by Analysis
• Material• Tolerance• Variability of the body/target environment• Atypical applications
Validation of Model by Test
• Analysis of tensile test to confirm ability to predict material behavior
• Validation tests for stents might include:- Flat plate loading- Radial expansion- Radial compression
Example:Flat Plate Loading Using Contact
Note: This “pinching” loading mode is distinct from “radial” loading
Are the Assumptions Satisfied?
Make adjustments/corrections as needed so that the model is predictive of the test
Additional Information and Insight From Analysis
Get information not available from device testing aloneInternal conditions: stress levels, degree of plasticity, residual stress, transformation fraction
Balloon Expandable Stent
1. Basic steps:Roll-down for catheter insertionInflation and DeploymentCyclic pulsation loading
2. Fatigue testing of full device to FDA required 400M cycles is a long process
Fatigue and Life Testing
• Long test times for full device• Reduce testing of multiple design iterations• Get insight more quickly• Need both analysis and testing
Cyclic Testing of Sub-specimen1. Before fatigue testing full device, get more
information in less time with sub-specimenHigher loading frequency, reduced test timeCycle to failure for a range of loadsDevelop part-specific S/N data
2. Extend with analysis, develop and interpret test conditions in terms of stress & strain
3. Make predictions for full device
Stent Segment
Sub-specimen
Stent Segment and Sub-specimen
Parnell, (2000)
Stent Segment Sub-specimen
Material Testing: Elastic/Plastic
• Need more detail than basic data from manufacturer (for example, Min. Yield, Ultimate, Elongation)
• Elongation is sensitive to the gage length tested• Reduction of area very useful, particularly for highly
ductile materials• Need full stress/strain curve with additional data like
reduction of area
Tensile Response of Elastic/Plastic MaterialE’
E
D
CB
A
0 ε
σ
Proportional limit
Yield stress
Ultimate stress
Linear Plastic Strain hardening
Significant necking
True Stress
Eng. Stress
Anderson (2002), Biomaterials
Typical stress/strain curve for steels. Strains become localized when necking occurs. Standard elongation highly dependent on gage length. Measured area reduction gives correct local strain.
Shape Memory Material (SMA) Applications
• Unique characteristics• Large recoverable strain range• Super elastic vs. Shape Memory (thermally
activated)• Self-expanding devices• Conditions after partial unloading• Load predictions
Applications for Shape Memory Alloys Materials that return to some shape upon appropriate temperature change
Applications:
Medical
IndustrialApplications
Home Appliance
Accessories clothing
Sports
Communi-cation
ShapeMemory
Shape Memory Material Properties
• DSC to determine transformation temperatures• Tensile test• Behavior as function of temperature • Super elastic material behavior
- General features (T > Af )- Stress-induced martensite and reverse
• Shape memory (reverting to learned shape)
NiTi Response to Temperature
T< Ms Shape Memory(residual strain recovered by heating)
Ms <T< Af Shape Memory(residual strain recovered by heating)
Af <T<Tc Superelastic (SIM)(full strain recovery)
T>Tc Plasticity before SIM(permanent residual strain)
As [K] Af [K] Ms [K] Mf [K]
188 221 190 128
Transformation Temperatures
Miyazaki, et.al., (1981)
Variation of SMA Structures
Pseudo-elastic behavior of SMATemperature induced phase transformation
Pseudo-elastic Stress-Strain Behavior
Material Testing: Shape Memory Alloy
• Transformation temperatures (DSC or other)• Stress/strain tensile curve with unloading• Application may require tensile data at additional
temperatures
Temperature Dependent Material Behavior of Shape Memory Alloys
NiTi Stent
Nickel-Titanium alloys show temperature dependent material behavior. Shape memory effect (that deformed specimens, regained their original shape after a loading cycle) is observed at a certain temperature.
Input:ASSσ
ASfσ
SASσ
SAfσ
−σ ,ASs
Lε To Cm Cs
ASfσ
SAfσ
−σ ,ASs
ASSσ
SASσ
Lε
To
Cm
Cs
ASSσ
ASfσ
SASσ
SAfσ
Input data for Mechanical SMA
Differential Scanning Calorimetry (DSC)
Shaw & Kyriakides, (1995), (courtesy of M.-H. Wu )
• DSC can be used to determine transformation temperatures of shape memory materials
• Heating curve: As,Af
• Cooling curve: Ms,Mf
• Austenite is Cubic (BCC)• Martensite is Monoclinic
Shape Memory Effect (SME)• Shape memory effect is a consequence of a crystallographically
reversible solid-solid phase transformation occurring in particular metal alloys (Ni – Ti, Cu based alloys).
• This transition occurs between a crystallographically more-ordered phase (called austenite) and a crystallographically less-ordered phase (martensite).
Stability for Martensite and Austenite Phases
Vulnerable Plaque
• Morphology• Tissue characteristics• Tissue properties and geometry become
important in evaluating device
Christensen, (2002)
Inverse Analysis Problem
• Correlate material properties to measured behavior
• Use to estimate ranges of properties for tissue• Example: estimation of vessel wall cyclic strains
from cine PC-MRI data (Draney, et.al., 2002)
Conclusions1. Testing and analysis are complementary;
both are essential2. Use together for maximum benefit
Reduce number of physical prototypesShorten development cycleAvoid surprises and delays
3. Applicable in all fields: ElectricalMechanicalBiomedical