Soft Tissue Material Properties and Viscoelastic Material Models for Finite Element Applications

Abstract

This thesis treats mechanical testing and modeling of eye tissue using the porcine eye as the representative model. The goal of the work is to combine finite element (FE) techniques and material data to generate an improved understanding of the mechanical behaviors of sclera with a reliable FE model. The mechanical properties of porcine sclera dogbone-shaped samples are studied in tension at various loading rates. A randomly selected group of samples is treated with a chemical cross-linking agent (genipin) to artificially stiffen their mechanical response. Response data is collected during both loading and relaxation of tissue samples at three quasi-static strain rates, 0.0006/s, 0.013/s, and 0.13/s. Measures of the tissue responses are then extracted and used to investigate average material property differences in treated and untreated tissue samples. The mechanical properties of porcine sclera are further investigated using whole, trimmed globes in a specialized testing system. The whole globes are perfused using an aqueous humor analog to generate loading conditions approximating those in vivo. The collected mechanical data is used to inform a series of material models in FE simulations of the mechanical tests. The method of generating material parameters is outlined, and a series of parameter sets are presented for each strain rate and each material treatment. The optimized parameter results are compared to the mechanical testing data using FE models of the mechanical testing process. Convergence of different FE models with increasing mesh density is demonstrated. Finally, FE models of the globe inflation process are informed with the optimized material model parameter sets and loaded to simulate the response of the globe under the action of intraocular pressure. Primary findings demonstrate a significant difference in material properties characterizing loading and relaxation of the genipin-treated and untreated sample groups. The parameter sets generated for each average group of samples are shown to fit the data well. Simulated globe inflation strain results were found to be greater in the untreated case than in the treated case. The untreated model strains compare well to photoacoustic image-measured strains and to strains reported in literature for similar processes.