Introduction to Durability Testing y g

What is Durability?
Ability of a vehicle to survive an expected service life Reliability is the probability of a products product s successful performance of an intended function up to a pre-determined life. Q lit i th consistency with which a population Quality is the i t ith hi h l ti of products perform throughout their life. Durability is a subjective term relating to the useful life of a product and is the term given to the type of testing used to determine the objective measures mentioned above. What are the major durability issues: Oft relate to reliability issues with components Often l t t li bilit i ith t Less often are structural

What is Durability?

Durability How long will it last ? What is the fatigue life? Repetitive loading emphasis Performance Does it behave as expected ? What are the characteristics ? Measurement emphasis

Durability Specifications
Durability is the ability to survive an expected service life Warranty period such as 3 years or 50,000 miles 150,000 miles, 6000 hours of targeted customer usage Safety Critical parts 3-4 targeted customer lifetimes

Why Do We Need To Test?

Test Topics
Next we will go on to discuss the: why, h who, when when, what, where where, & how of testing testing.

Why Do We Test?
Confirm physical performance meets predicted design targets g g Ensure correct operation throughout the service life of our products d t y g y Identify and correct manufacturing and assembly incidents prior to final vehicle release Provide complete durability coverage complete Significant recall and customer p g perception costs p associated with field failure

Why Do We Test?
The variability of end-users, manufacturing processes, materials, etc. processes materials etc can severely impact durability
Loads L d

Durability D bilit
Material ls

Geometry G t

To efficiently address durability issues we must first understand and replicate the durability p p y phenomenon

Why Do We Test?
Consider the impact of a component failure: Ignition switch failure ($15 part) $15 (part) + $35 (labor) * 8M (vehicles) = $400M Recall Notice $0.39 Postage stamp * 8M (vehicles) = $3 1 Million $3.1 Million Just for the Stamps!

Who Tests?
Vehicle Manufacturers (OEMs) Tier 1 suppliers (suppliers to OEMs) Tier 2 suppliers (suppliers to Tier 1) Independent test houses Government agencies

Who Tests?
Vehicle manufacturers (OEMs) passing test responsibility to Suppliers p y pp Testing required to confirm manufactured components and sub-assemblies meet OEM t d b bli t performance goals Suppliers increasingly have to accept consequential liability for non-performance of products in service More testing on components and sub-systems, less t ti on full vehicles. l testing f ll hi l

Who Tests?
Suppliers need to test both for validation of product p performance and to p protect against warranty liability g y y Testing to OEM requirements will show specification conformance ifi ti f Testing is necessary to support q g y pp quality y improvements and cost reduction activities Component reliability evaluated through testing

When Do We Need To Test?

Committed $ d
Tooling C o n c e p t Sub-system Manufacture Design Vehicle Manufacture

P r o d u c t i o n

1. Design Confirmation 2. Sub-system Validation 3. Product Validation 4. Quality 4 Q lit

Where Do We test?

Where Do We Test?
Proving Ground Testing Full vehicle testing on the Proving Ground G d ~$3/mile 3 months Functional running vehicle required

Where Do We Test?
Proving Ground Activity (Customer Correlation) Pave or Belgian Block Washboard Pot Holes Curb Strikes Rough Road Cross Country Ride & Handling Other Special Events

Where Do We Test?
Laboratory Testing Full vehicle testing in the Laboratory ~$0.70/mile 3 weeks Functional running vehicle NOT required

Where Do We Test?
Laboratory testing moves into the Computer Laboratory and Proving Ground Testing is testing of physical prototypes, sub-systems and components Computer methods are performance evaluation of virtual prototypes using analytical testing

Where Do We Test?
Virtual Tools commonly employed today: Finite Element Analysis Life Predication Multi Body Dynamics

Analytical = Prediction Physical = Confirmation

What Do We Test?
Durability is not the only Vehicle attribute, testing is also conducted for: Crash Powertrain NVH Ride and Handling Testing occurs at the component, sub-system, and full vehicle levels

What Do We Test?
Full Vehicle

System

or Component

What Do We Expect From Test?

Functional Evaluation Confirmation of the basic kinematics of the specimen (e.g. suspension, window winding mechanism) Parameter Evaluation Measurement of the characteristics of the specimen (e.g. roll steer, modal survey) Durability Evaluation Confirmation that the wear and fatigue properties of the specimen meet design goals

What Do We Test What Are The Objectives?

Target mismatch Design Engineer Targets Determine average service load then keep all stresses below 2/3 of Yield keep maximum identified stress using Finite Element analysis below fatigue limit Test Engineer Targets With 99% confidence assure that less than 5 parts in 100 will fail in service Service is defined as 95th percentile customer, 160,000km usage Does meeting the design targets necessarily mean meeting the test targets?

What Do We Test What Are The Objectives?

For each durability test we must define failure. A few possible definitions of failure:
- Fracture of the specimen into two or more pieces. - Development of visible cracks of a certain size. - Large-scale yielding of the specimen. - Ch Change iin stiffness of th specimen. tiff f the i - Excessive deflection of the specimen. - Excessive wear of the specimen specimen.

Design tests so that failure is detectable. g

What Do We Need to Achieve?

Safety Factor P(x)

Service loading

Strength

What Do We Need to Achieve?

Safety Factor P(x)

Service loading Over-design

Strength

What Do We Need to Achieve?

Safety Factor P(x)

Service loading Under-design

Strength

What Do We Include in Test What Are The Environmental Considerations?

Mechanical Loading Forces applied to the specimen and reacted either inertially or though a fixed load path Thermal Load and Humidity Thermal environment, ambient, under-hood, sun load and humidity may need to be applied if specimen durability is sensitive to these effects (e.g. thermoplastics) Corrosion and Abrasion Salt spray may be needed to correctly evaluate durability for ferrous materials Dust may need to be introduced to correctly evaluate wear

What is Fatigue?
Fatigue of a metal is a p process in which the metal experiences progressive structural damage from repeated cyclic loading. Fatigue results in the formation and growth of f ti d th f cracks. The cracks may propagate to cause complete fracture of the component.

What is Fatigue?
Fatigue life is a measure of the duration of cyclic loading required to form and g g q grow cracks to a predetermined size.

Log (Alternating Stress) Log (a)

103

104

105

Fatigue life consists of two principle stages: Crack initiation Crack propagation

106 107 108 Life to Failure (Cycles)

Test Loading Methodologies

Descending order of realism and time Customer Usage Proving Ground Road Simulation Shaped Random Noise End Level Sequence End-Level Block Cycle Constant Amplitude The challenge to test engineering is to achieve an acceptable level of realism in the time allotted to be economically competitive!

How Do We Test?

Durability Testing Approaches

Success Testing (no failures occur)

Failure Testing

L Large sample size l i required to achieve confidence Little understanding of product b h i d t behavior

Smaller sample size (more practical). Understand product weaknesses (make continual improvements). Ensure that lab tests duplicate field failure modes.

How Do We Test?

Success Testing

Example Field Testing:

Assess reliability by testing numerous p y y g parts up to p one design life, with no failures.

Downfalls:
1. Sample size:
How many samples must be tested without failure to have 90% confidence of 98% reliability?
1 n +1

0.98 = (1 0.90) n = 113 parts!!! (YOU tell the boss! )

2. No information on product behavior, weaknesses, failure modes modes. R = Estimate of reliability of population.
R = (1 C )
1 n +1

C= n=

Confidence (I.e., probability that the population reliability is at least as high as the estimate, R). Number of specimens tested.

How Do We Test?

Failure Testing

If failures occur, can estimate life distribution using Weibull distribution.

Now, reliability for any time, t, can be estimated. Ideally, want 6-10 samples tested to failure, but can estimate Weibull parameters with fewer samples (l t ith f l (lower confidence)
Unre eliability (Wei ibull Probabil Scale) lity

Time to Failure (log scale)

F (t ) = (1 e)

F(t) = = =

fraction that fails by time t (Unreliability) shape parameter characteristic life

Test Types
Laboratory Durability Test Acceleration The goal of testing i to t ti is t induce fatigue damage faster than i th in service i (or on the proving ground)

Test Types
Accelerating accumulation of damage achieved by: Increasing amplitude Increasing frequency Removing non-damaging content Main approaches Cyclic Block Cycle Shaped Random Noise End Level Sequence Simulation

Accelerated Durability Testing Simple Cyclic

Measured Service Strain

Range-Mean g Rainflow Matrix m r

Material Properties

Calculate fraction of fatigue life used Adjust applied load to achieve selected strain level Cyclic Test Load S t

N Select strain level and cycle count d l

Accelerated Durability Testing Simple Cyclic

Advantages Test Acceleration (1000 x real time) due to increased average amplitude of applied load and increased rate of strain cycle application Disadvantages Multi-axial loading not p g possible Dynamic load amplification (frequency effects) due to specimen resonance not replicated Test results may be a poor indicator of service life if the material properties, manufacturing process or design of the specimen changes Not suitable for inertially-reacted systems.

Accelerated Durability Testing Block Cycle

Measured Service Strain

Range-Mean g Rainflow Matrix m r

Material Properties 0 N

0 r

t Develop equivalent block cycle strain history Block cycle Adjust applied loads to achieve selected strain levels test load S t

Remove strain ranges below fatigue limit

Accelerated Durability Testing Block Cycle

Advantages Test Acceleration (100 x real time) due to removal of non-damaging cycles and increased rate of strain cycle application R l ti l i Relatively insensitive t material and iti to t i l d manufacturing process changes Multi-axial loading possible Disadvantages True multi-axial loading not possible multi axial Dynamic load amplification (frequency effects) due to specimen resonance not replicated Not suitable for inertially-reacted systems.

Accelerated Durability Testing End-Level Sequence Loading q g

Description Sequence of end-levels end levels
Amplitude t

Applications
Specimens subjected to variable amplitude loading (uniaxial and multiaxial). When sequence effects are important important. When phase relationship between channels must be maintained. Fi d Fixed-reaction component testing. i i

Accelerated Durability Testing End-Level Sequence Loading q g

Advantages Reproduces realistic sequence of amplitudes amplitudes. Allows simulation of sequence effects. Maintains approximately correct phasing between channels. Allows frequency acceleration (fixed-reaction tests). ) Disadvantages Some errors in phase simulation in multi-axial tests. Not as easy to accelerate as block cycle tests. y y Not suitable for inertially-reacted systems.

Accelerated Durability Testing Shaped Random Noise Loading p g

Amplitud de

Amplitude

f
Description Random signal with specified spectral shape. Applications Specimens subjected to variable amplitude loading (uniaxial or uncorrelated multiaxial). Single axis vibration tests. tests Squeak and rattle

Accelerated Durability Testing Shaped Random Noise Loading p g

Advantages Easy to specify. specify Minimal specification and storage. Replicates a wide variety of natural vibrations. Disadvantages No control of end levels. levels Uncorrelated multi-channel tests may not induce realistic fatigue damage. Does not well-represent periodic overload conditions and their associated sequence effects.

Accelerated Durability Testing Road Simulation

Measured Service Strains and Loads (yn) 0 N xn yn t t Simulation Loads Desired Responses Material Properties Measure rig transfer function Hxy between drives xn and responses yn Remote Parameter C R P Control l (RPC) process compensates for multi-axial dynamic test system response

yn

Accelerated Durability Testing Response Simulation p

Advantages Test Acceleration (10 x real time) due to removal of non-damaging events Multi-axial loading effects correctly replicated Dynamic loading effects correctly replicated Minimal modification of real road data Correct sequence of loading reproduced Accurate replication of failure mode and failure location Disadvantages Currently requires instrumentation and measured loads data for different test specimens (e.g. different vehicles)

Challenges with Durability Assessment

Validation of durability targets occurs late in the development p g p program The inherent lack of confidence in current predictive tools t l necessitates physical t t it t h i l test Most automotive components under g multi-axial p go variable amplitude loading

Global Durability Trends

Testing is expensive but testing is necessary Lack of experience can result in expensive mistakes Increased focus on CAE generates greater demand for component data and validation tests Increased reuse of existing test data requires test focused data management

MTS Systems Corporation

Global Market leader in testing systems and software in Durability and Ride & Handling. y g
6DOF Long Stroke Road Simulator High Frequency MAST

SWIFT Wheel Force Transducer

MTS Systems Corporation

Broad Testing Solutions

MTS Systems Corporation

Durability Test Type MTS Product Line
4-Poster / Spindle Coupled Simulator Drivetrain MAST Damper / Sh k & D Shock Elastomer , Ball Joint, Exhaust & Steer TestStand & Testline

Full Vehicle

System

Component

RPC & MPT Software

4-Poster / Spindle Coupled Simulators

Full Vehicle System Component

MTS Model 320 Tire Coupled Road Simulator (Four Poster) Fast setup time Able to produce results quickly. Actuators can be mounted below floor level - Drive-on simulators - Able to interface with environmental chambers

MTS Model 329 Spindle-Coupled Road Simulator U t 6 DOF Control at each wheel Up to C t l t h h l Highest degree of simulation accuracy available Configurations available to reproduce - Body roll (slalom events) - Body p y pitch ( (brake events) )

MAST Systems

Full Vehicle System Component

MTS MAST (Multi-Axial Simulation Table) Systems Accelerated replication of service data for - Body components and assemblies - Seats - Fuel Tanks - Radiators - Engine mounts - Exhaust systems - and more!

Model 353.20 High Frequency MAST Controlled frequency exceeding 100 Hz 2 meter table Compact design Applications include - Seats - R di t Radiators - Instrument Panels - HVAC Systems

Drivetrain Systems

Full Vehicle System Component

MTS Model 814 Spinning Torsion System Designed for torsional dampers in - Manual clutches - Torque converters Durability testing of the damper under - Rotation and - Torsional solicitation Hysteresis measurements at multiple speeds

MTS Shaft Test Systems D bilit T ti on Durability Testing - Driveshafts - Halfshafts - Joints Proven tools for simulation

Shock, Damper & Elastomer Systems

Full Vehicle System Component

MTS Model 835, 849, 850 and 851 systems Performance characterization & durability testing for - Shock absorbers - Dampers - Struts

MTS Models 831 and 833 Systems Ch Characterization and durability t ti t i ti d d bilit testing 100 to 400 Hz typical High Frequency system .01 to 1000 Hz

Ball Joint, Exhaust and Steer Systems

Full Vehicle System Component

MTS Model 324 Exhaust Test Systems Valves to tailpipe Full Simulation Testing B Based on proven MTS MAST T h l d Technology

MTS Steering Test Systems S Several systems t choose f l t to h from including i l di - Model 335 & 337 Rack & Pinion Systems Lateral, Vertical and Steering Loads Block cycle or Time History Reproduction

TestStand & Testline Systems

Full Vehicle System Component

TestLine For the Do It Yourselfer Versatile and reliable array of test components C t Effective Cost Eff ti

TestStand Solutions T t system design and consulting services Test t d i d lti i Acquire leading-edge technology - In a cost-effective manner Collaborative process to augments in-house skills.

RPC & MPT Software

Full Vehicle System Component

RPC Pro Software (Remote Parameter Control) Full-featured Road Simulation Advances signal processing and editing F ti Fatigue Analysis A l i State of the art simulation technologies

MPT (Multi-Purpose TestWare) Software E Easy to use t For creating and automating many test types Test Procedures can be built that include - Function Generation - Data Acquisition q - Event Action - Triggers

MTS Systems Corporation

MTS Simulation Test Consultants Instrumentation and Data Acquisition Analysis and Editing Fatigue Correlation Test Development Fixture Design Global Support Training

Engineering a better world for ground vehicles