Computational and experimental characterization of the mechanical behaviour of a novel corneal implantable device

Keratoconus is a non-inflammatory eye disease that causes progressive thinning of the stromal cornea and collagen weakening resulting in a conical corneal shape leading to irregular astigmatism. All currently available treatments do not guarantee predictable outcomes. The GROSSO® implant is a corneal device made of Nitinol alloy, designed to reshape the whole cornea by virtue of a dome-shaped design that follows the physiological curvature. The aim of this thesis is to investigate the mechanical behaviour of the GROSSO® implant through computational simulations performed with Finite Element (FE) Analysis. The investigation focused on two critical situations. The first replicates a crucial phase of the surgical procedure inside the patient’s intracorneal pocket consisting in a 90° folding of the opposite edges of the device (“bending” simulation). The second replicates a crush test of the device to evaluate its response to compression between two plates (“crushing” simulation). A 1D and 3D model of both Version 4 and 5 of the GROSSO® implant were developed from 2D geometry provided by Recornea, using Rhinoceros (Robert McNeel & Associates, USA). The models were then meshed using beam and tetrahedral elements, for the 1D and 3D models respectively, in HyperMesh (Altair Engineering, USA). The displacement-driven simulations were conducted in ABAQUS Standard (Dassault Systemes Simulia Corp., USA) and a sensitivity analysis was carried out gradually reducing the size of mesh elements. In addition, three different coefficients of friction were set in the crushing simulations. Data of Maximum Principal Strain (MPS) and Reaction Forces were then extracted. A previous characterization of the Nitinol alloy indicated that at MPS values higher than 8% plastic deformations were present. Thus, the MPS extracted from FE Analysis was compared with the value of 8% to confirm whether the device operates under elastic conditions. In the bending simulation the MPS reached was 6.13% for Version 4 and 6.01% for Version 5 respectively, while the maximum force reached was 0.08 N. The MPS reached in the crushing simulation with a friction coefficient of 0, 0.1, 0.57 was respectively: 2.89%, 2.86%, 2.66% for Version 4 and 2.56%, 2.51%, 2.46% for Version 5. The bending was then experimentally tested both at the Politecnico di Torino and at the School of Life Sciences FHNW in Switzerland. At the Politecnico di Torino, the ElectroForce Planar Biaxial TestBench instrument (TA inst.) equipped with a load cell of 225 N was employed. A nylon suture thread was wrapped around two symmetrical points on the outer ring, and its ends were secured to the pulleys. A single displacement of 8.75 mm at a speed of 0.5 mm/s for each pulley was set. The compression cycle was repeated 5 times in a row in a Version 4 physical device. The maximum force reached was 0.44 N, in disagreement with the computational result since the force range was too low for the sensitivity of the implemented load cell. At FHNW, 3D clamping devices were specially made to clamp the implant and then mounted on the testing machine Bruker Tribolab with a load cell of 22 N. The device was compressed and extended 5 times in a row. Both compression and extension forces were measured. The maximum force reached was 0.09 N. An excellent agreement between the compression-extension cycle of the experimental test performed at FHNW and the numerical one was found. The study could be further extended including the experimental validation of the crush.