Fatigue Analysis of Nitinol-Based Mechanisms

Figure 1: Equivalent Total Strain of stent segment under systolic pressure.

Introduction

Fatigue analysis of Nitinol-based mechanisms presents unique challenges compared to similar analyses performed on non-superelastic alloys. While standard metal fatigue can be investigated using S-N curves and stress-based Goodman diagrams, a more accurate method for nitinol uses strain measurements. This report details a sample investigation of the fatigue life of a nitinol-based stent.

Fatigue in Nitinol

Fatigue behavior in Nitinol is a fast-developing science with a 2012 review of its behavior documented in "Mechanical fatigue and fracture of Nitinol" [1]. While testing methodologies are similar to other metals, strain has been shown to be more predictive of failure than stress. As stents are a primary application of nitinol, much of the fatigue testing has been performed on stent subcomponents or whole stents. This testing is commonly performed with non-zero mean stresses to mimic in vivo conditions. Testing with results illustrated in Figure 2 shows that under a mean strain value of ~6.5%, fatigue life is above 10^7 cycles when the strain amplitude is below0.4%. This diagram is called a constant life diagram. 

Fatigue life in nitinol can change drastically based on temperature, and the constant life diagram assumes that the material begins in an all austinite phase. Other factors that affect fatigue life are similar to other metals, e.g., surface finish. Most experiments are performed on electropolished and heat-treated materials similar to cardiovascular stents. As such, a manufactured stent will behave similarly to a constant stress diagram if processed following best practices.

Figure 2: Strain Amplitude vs. Mean Strain of Nitinol stent subcomponents tested to a 10e7 cycle fracture [1]

Analysis

The sample analysis is performed on a subsection of a stent. The stent is compressed to an insertion diameter and then released to come into contact with a blood vessel. The blood vessel is approximated as a homogeneous cylinder, a more detailed blood vessel model and stent deployment is described in my report on Stent Mechanical Simulation. In addition to the contact forces of the stent, the blood vessel is pre-strained in the axial direction, and systolic/diastolic pressures are applied to the inner surface. 

Stress Amplitude is found by subtracting the strain at diastolic and systolic pressures and then dividing this quantity by 2. Mean strain is found by adding instead of subtracting

Results

The maximum mean strain in the stent is 6.5% (Figure 4), which indicates that the stent is approaching the region of the constant life diagram where fracture can occur before 10^7 cycles with less than 0.4% (Figure 3) strain amplitude. Thankfully, the strain amplitude of the stent is 0.1%, placing the fatigue life well inside the region with over 10^7 cycles. This fatigue life is considered to be sufficient for cardiac stents

Weaknesses and Future Work

The fatigue analysis documented here only considers the fatigue life due to pulsatile pressure-induced deformation. However, stents can experience other sources of deformation. All stents will experience large strains due to crimping for implantation, the stent of this study experiences 8% strain during crimping. This single cycle of large deformation is unlikely to affect the fatigue life significantly. However, other deformations can occur in peripheral stents (e.g. femoral artery stents) due to body movement, sitting, etc. Fatigue life including these loads can be performed when estimations of their induced strains and occurrence frequencies are calculated. With that data, a cumulative damage fatigue model can be created to estimate the fatigue life more accurately. 

Figure 3: Strain amplitude

Figure 4: Mean strain

References