I am a research assistant in the Computational Mechanics laboratory in the Department of Civil and Environmental Engineering. My main area of research interest involves investigating failure mechanisms in materials. For my MS project, I was working with an elastomeric material which was obtained from "Rubber-Cal". The material I used was a sheet rubber called hard neoprene. It's a blend of Styrene Butadiene Rubber, Neoprene, and Nitrile Rubbers and is a very typical material used for industrial gaskets, vehicle bumpers, pads, seals, O-rings, etc. Failures in these types of materials can cause catostrophic situations. Examples of some notable elastomer failures are the Firestone Wilderness tire failures and the Challenger space shuttle's O-ring failure. The study of this elastomer's failure poses some difficult experimental and analytical issues.
The experiments I conducted were standard tension tests, stress-relaxation tests, fracture or single edge-notch tension tests, and three-point bending tests. These experiments were conducted as nearly as possible to ASTM standards, when applicable. Even though the ASTM standards are straight forward in specifying the means of testing plastic materials, the material properties of the elastomer I have been testing make it very difficult to test exactly according to the standards. Most of the problems I have experience come about during fracture tests.
The elastomer is very flexible and during pure tension tests, it exhibits a total percent strain of approximately 300%. During fracture testing, the material experiences blunting almost immediately upon load initiation. This makes it practically impossible to ensure fast fracture. The flexibile nature of the elastomer also makes compact tension (CT) tests very unreliable and unrealistic. As a specimen is loaded, it shows large stress concentrations and deformation in the area surrounding the pin holes for loading and also incurs significant out of plane warping. Using a standard CT test specimen makes it difficult to ensure mode I loading and failure most often occurs in a plane-stress or mixed-mode state as opposed to plane strain.
I abandoned CT test specimens and opted to perform fracture tests on a rectangular specimen that could easily be loaded to in a manner to simulate a far-field stress state. The photo above was taken during one of my experiments. The specimen tested was a regular dumbbell shaped specimen with a sharp crack initiated by a razor blade. The video that will soon be below is a video that I made of one of my fracture experiments. Notice how quickly the sharp crack tip blunts itself during loading. You can also see that the simulation video in numerical results section is a nice representation of the actual experiment.
One of the most interesting, difficult, and rewarding aspects of my research has been my attempt to model the failure of my specimens via the finite element method (FEM). I have been working on implementing the FEM, in MatLab, to this highly non-linear problem. I have successfully implemented a compressible Neo-Hookean constitutive law, which I have used with the free meshing software, G-mesh, to model a variety of tensile test specimens. The most notable of which were a dumbbell specimen similar to my tension specimens and a fracture speciment with a sharp crack tip that opens as the specimen in loaded. I have successfully implemented a method of allowing the crack to propagate as the specimen is loaded as well. My hope is that I will soon be able to successfully model my fracture tests completely to the point of failure. The video below is a simulation video created with PoVRay to simulate fracture tests.
I have found that for my numerical models to offer a good representation of the load versus displacement curves, that I have obtained through experiments, it is necessary to model a nearly incompressible Neo-Hookean material. The reason is that the Poisson ratio required to offer a good graphical representation of the experimental data in simulation is very close to 0.50. I have recently implemented the nearly incompressible form for a Neo-Hookean material and I am currently trying to successfully match the material parameters.