Abstract
Nitinol is a shape memory alloy that has become very popular for medical applications. Specifically, Nitinol tubing is used as peripheral stents, cardio stents and other medical implant devices. Medical implants subjected to blood pressure experience approximately 40 million cycles per year, which necessitates a life expectancy of at least 600 million cycles. Consequently, modeling the fatigue life of Nitinol is critical because failure could result in severe consequences, if not death. The purpose of this work is to model a rather robust set of fatigue data for Nitinol tubes. A phenomenological distribution function for fatigue life is considered. The form of the distribution function includes a kernel that incorporates time dependent loading, mechanical breakdown, and statistical behavior. The approach is a form of the classical accelerated life model in which covariates are included to account for physical attributes that directly influence lifetime. Because of the generality in the formalism, the model is applicable for a variety of loading conditions. The modeling for the Nitinol data is a combination of traditional stress–life methods with a Weibull distribution function. The proposed approach incorporates stress dependencies in the distribution parameters. Also, the Weibull distribution is assumed to be in the form of a three–parameter distribution rather than the more frequently used two–parameter. To assess the validity of the proposed methodology confidence bounds will be estimated for the data.
Higher Moment Statistics for Fitting Random Fatigue Life Data to Weibull Distribution Parameters
