Worked with a sponsor, Hollister, to create a computational model of their physical prototype of a stoma. The physical prototype had various limitations: not easily adjustable to different stoma types, very time-consuming to produce, and wasteful. 

The prototype was rebuilt in SolidWorks and imported to ANSYS. Tensile testing was done on the individual layers of the prototype to calibrate the material model in ANSYS. Further mechanical testing, like compression and 3 point bending, was done on the whole prototype to validate the computational model.

The computational model is easily adjustable to change material model parameters and test unique stoma topographies.

Waller et al. created a finite element model on the effect of baseplate convexity on stoma site [6]:

• Constructed a 3D model of muscle, fat, skin, and stoma using material properties found in other research

• FEA simulation visualized distribution of strain 

Limitations:

• FEA model assumed homogeneous material properties and simplified geometries

• Flat abdomen simulation

• Only one type of force application was analyzed

• Not validated with real-world data

A decision matrix was made to determine the mechanical tests to be conducted on the Hollister prototype to derive material properties.

The criteria included how relevant the test was to the actual mechanics occurring at the stoma site, how accessible the equipment needed to conduct the tests were to us, and the computing power required to replicate that test on ANSYS.

Tensile testing and torsion testing on the individual layers were chosen for material model calibration while compression and 3 point bend testing were chosen for the computational model validation.

Design matrix evaluating different types of mechanical testing

The physical prototype consists of three main layers, each with unique material properties. Tensile testing was conducted on each layer, three samples per four different speeds until failure, to determine if the material had viscoelastic behaviors. This testing was done according to ASTM D412-16 (2021) standard, with the layers being prepared into straight section specimen (1.5 by 3 inches).

Using MATLAB, an ANOVA test was done on the stress-strain curves collected from tensile testing to determine if the material was viscoelastic. This resulted in the stress-strain curves of different speeds being significantly similar, so the material was not considered viscoelastic.

Torsion testing was also conducted on the same specimen as tensile testing. However, the machine used did not have enough precision to be useful for material model calibration, and ideally, for torsion, a cylindrical sample of the material would be used. So, the data from this testing was excluded from calibration.