|
Track 1 Session 6
10:20 to 11:30 a.m. Thursday
April 12, 2007
Basic Methods and Guidelines for
Reliability Validation of Automotive Components at High Mileage
Reliability tests realized on a
limited number of vehicles or components cannot guarantee that the
cumulative failure rate will be lower than 1% at a given high
mileage, which is not an ambitious reliability objective. A more
efficient approach consists in conducting accelerated tests,
combined eventually with numerical simulations. This approach raises
a difficulty: how to forecast the future reliability of the
component in the field from test/simulation results? The assessment
method must take into account two important sources of variability:
the manufacturing scatter and the variability of customer use. We
present four basic methods that enable the analyst to establish the
relationship between field reliability and test/simulation results.
Due to its wide range of applications, the practice of the
Stress-Strength method is particularly detailed. The implementation
of these methods is illustrated by several applications on power
train components. Their limits and success conditions are discussed.
From the experience acquired, we deduce some guidelines for the
design and the interpretation of reliability validation tests.
Key Words: Experimental Reliability, Reliability
Assessment, Reliability of Automotive Components, Stress-Strength
Method, Degradation Modelisation for Reliability Assessment
Paul Schimmerling
Renault
Rueil Malmaison, France |
 |
|
