Uncertainty Identification of Damage Growth Parameters using Health Monitoring Data and Nonlinear Regression

Alexandra Coppe, Raphael T. Haftka, and Nam-Ho Kim
Submission Type: 
Full Paper
Supporting Agencies (optional): 
NASA and Air Force
AttachmentSizeTimestamp
phmc_10_074.pdf441.25 KBOctober 12, 2010 - 9:39am

When it comes to identifying model parameters such as damage growth parameters in Paris law for example, Bayesian inference is a popular method. However, it involves substantial computational cost, especially with increasing number of parameters. When the prior distribution for the parameters is not narrow, non-linear regression may provide almost all the benefits of Bayesian updating at a small fraction of the computational cost. In this paper we apply this approach to the identification of damage growth parameters. As a first step we simplify the problem to a single parameter in order to compare it with the same problem solved using Bayesian inference. We then discuss the issues related to uncertainty quantification in the case of a highly non-linear problem.

Publication Control Number: 
074
Submission Keywords: 
prognosis
structural health monitoring
non-linear least square
Uncertainty Quantification
damage propagation
remaining useful life (RUL)
Submitted by: 
  
 
 
 

follow us

PHM Society on Facebook Follow PHM Society on Twitter PHM Society on LinkedIn PHM Society RSS News Feed