Cutting models for bio-composites

Abstract

On the soft composite front, the cutting of single ply-reinforced hydrogels is investigated in a sensor-rich testbed comprising of both a load-cell, a submerged hydrophone, and a digital camera. The mapping between the pressure and force signals reveals four distinct regions of the cutting process, viz., gel indentation, steady state gel cutting, ply indentation and ply failure. The force and pressure signals are complementary in that a burst in the pressure signal always corresponds to a drop in the axial thrust force. The hydrophone signal is used as the diagnostic input signal to model the time-dependent evolution of the cutting forces. The hybrid modeling framework relies on region-specific modeling strategies spanning analytical, finite element method and mechanistic thrust force modeling techniques. The model is first validated using single-ply experiments and then used for to predict the effect of ply-spacing in multi-ply composites. The findings show that the hybrid modeling framework is capable of predicting the temporal evolution of the thrust force, while also providing insights into the underlying mechanics of deformation seen in these soft composites. Overall, this work establishes the hydrophone pressure signal as an effective in-situ process monitoring sensor to be used during the cutting of fiber-reinforced hydrogels.

The research presented in this thesis aims to investigate the cutting behavior of hard and soft bio-composites, specifically, bovine cortical bone and fiber-reinforced hydrogels respectively. For bovine cortical bone, a microstructure-based finite element model to simulate fracture cutting of bovine cortical bone is developed. For soft composites, investigation is done into the experimental and modeling aspects associated with the cutting of fiber-reinforced hydrogels, with a focus on the application of the hydrophone as a process-monitoring sensor.On the bovine cortical bone front, fracture cutting of both the haversian and plexiform components of bovine cortical bone is successfully simulated using a cohesive zone-based approach. This model presents an improvement over existing bone cutting models in this category that have relied only on parametric inputs for the material properties. A key novelty of this work lies in the fact that the cohesive zone parameters associated with each of the failure modes are heuristically extracted using their specific acoustic emission signature. This approach ensures that the Cohesive Zone (CZ) parameters capture the specific failures observed in the cutting experiment. The cohesive tractions were first extracted using the experimental results for the +20° rake tool. They are then successfully used to predict the results for the 0° rake tool.

The process planning efficacy for model is also demonstrated using parametric studies spanning both the haversian and plexiform components. While the heuristic approach has its limitations given the property variations seen across different bone types and age, the maturation of such experimental data sets combined with machine learning techniques, can yield an effective mapping between the heuristically obtained CZ model parameters and the local microstructural property variations in bone. This avenue presents a key opportunity to study natural composites whose individual phases cannot be easily characterized.