Failure Modes Analysis of Fatigue S-N Test Data with Small Sample Size  
Author Zhigang Wei

 

Co-Author(s) Fulun Yang; Burt Lin; D. Gary Harlow

 

Abstract Multiple-modal statistical behavior is an important aspect in material durability applications. A variety of materials subjected to diverse loading conditions have exhibited bimodal or multiple damage and failure behaviors. Several innovative statistical modeling methodologies for bimodal and multi-modal damage growth and failure analyses have been developed recently based on either lognormal or Weibull distributions. However, these methods are usually applied to superalloys with a large amount of test data, and the application of the methods to test data of materials used in auto exhaust components with small sample sizes is essentially unknown. Tests of auto exhaust components and systems are usually expensive and, therefore, the test sample size is typically very limited in order to keep a relatively low budget. In addition, a linear S-N curve is often assumed for fatigue S-N test data in auto exhaust applications and any multiple-modal behavior is often ignored. In this paper, multiple failure modes in the fatigue data of auto exhaust systems is investigated first, then the goodness-of-fit of lognormal and Weibull distribution functions for these test data with both single and multiple failure modes are demonstrated and compared. The possible underlying physical basis of fatigue damage process is also analyzed and discussed.

 

Keywords Durability, Bimodal failure modes, Multiple failure modes, Sample size
   
    Article #:  18101
 
Proceedings of the 18th ISSAT International Conference on Reliability and Quality in Design
July 26-28, 2012 - Boston, Massachusetts, U.S.A.