Manufacturing of thermosetting polymers and composites using frontal polymerization: A numerical study

Frontal polymerization;Reaction-diffusion;Finite element, Computational multiphysics, Composites manufacturing

Goli, Elyas

关键词: Frontal polymerization;    Reaction-diffusion;    Finite element, Computational multiphysics, Composites manufacturing;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/107930/GOLI-DISSERTATION-2020.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Fiber-Reinforced Polymer Composite (FRPC) materials are vital to today's aerospace, automotive, marine, construction, and energy industries, and will be integral to the next generation of lightweight, energy-e cient structures owing to their excellent specific stiffness and strength, thermal stability and chemical resistance. Yet, the wider use of highperformance thermoset composites is limited by the complexity and cost of the fabrication process, which usually requires the monomer to be cured at high temperatures (around 180oC) for several hours under combined external pressure and internal vacuum. Curing is generally accomplished using large autoclaves or ovens that scale in size with the component. Hence, this traditional curing approach is slow, requires a large amount of energy, and involves substantial capital investment, resulting in an expensive, time- and energy-intensive process.Over the past few years, a new out-of-autoclave/out-of-oven manufacturing method for FRPCs based on Frontal Polymerization (FP) has been introduced by the Autonomous Materials Systems (AMS) group at the University of Illinois that greatly speeds up the process and substantially reduces the energy cost (by many orders of magnitude) duration of the manufacturing process. FP is a self-propagating reaction driven by exothermic polymerization, where an advancing front is formed by a local thermal stimulus applied to a solution of monomer and initiator. The heat generated by the exothermic reaction diffuses forward to advance the propagating front, resulting in a self-sustained process. To support the development of this new manufacturing method, we simulate the process by developing a nonlinear, transient, multiphysicsfinite element solver. The solver is used to (i) better understand the underlying physics, and (ii) to make a link between the input parameters such as the resin chemistry and the properties offibers, and the manufacturing process andfinal product. Thefirst part of the study focuses on the investigation of FP in a neat Dicyclopentadiene (DCPD) resin, in the absence offibers. In a continuum-level approach, we numerically solve a reaction-diffusion system of Partial Differential Equations (PDEs) to model the initiation and propagation of polymerization fronts. The second part of the study deals with the merging fronts and the resulting thermal spike at the merger. We shed some light on the physics behind the temperature overshoot where fronts meet and propose a simple framework to predict its magnitude. As a link between the fabrication of polymeric components and composites manufacturing, we introduce a conductive element into the resin to analyze how it affects the front characteristics. We then propose a homogenized thermo-chemical model to simulate the FP-based manufacturing of unidirectional carbon- or glass-fiber-reinforced DCPD-matrix composites.Afterwards, we solve the model using a nonlinear finite element solver and validate the results with experimental data. We demonstrate that the validated framework can be used to tailor the chemistry for specific composite applications. In most cases, the polymerization front propagates in a steady fashion. However, under some conditions, the front experiences instabilities, which do affect the quality of the manufactured composite part. Here we investigate whether these instabilities can be captured by a reaction-diffusion model in the absence of convection in the monomer ahead of the front. In particular, we use a coupled thermo-chemical model and an adaptive nonlinear finite element solver to simulate FP-driven instabilities in DCPD and in carbon-fiber DCPDmatrix composites. With the aid of 1-D transient simulations, we investigate how the initial temperature and the carbonfiber volume fraction affect the amplitude and wavelength of the thermal instabilities. We also extract the range of processing conditions for which the instabilities are predicted to appear.Investigating the effect of boundary conditions through convection and contact boundaries is the focus of next part of this thesis. In this line of work, we evaluate the effect of different parameters such as the film coefficient, resin volume, and diffusivity of the tooling plate on the front velocity and maximum temperature during the manufacturing of thermosetting polymers. The results of this study may serve as a guide map for chemists and manufacturing engineers to adjust their experimental and industrial setups.In the last part of this thesis, we integrate the frontal polymerization with machine learning to tailor the chemistry for a desired manufacturing strategy. To that effect, we first develop and implement a FeedForward Neural Network (NN) model in Python. We then train the model using the input data generated with normal distribution and corresponding output data extracted from the steady-state solver. Finally, in an `inverse approach', we utilize the model to predict the cure kinetics parameters for given front characteristics. The model will enable us to optimize the cure kinetics of the matrix phase of the composites for a specific application (a desired volume fraction of fibers) and a favorable manufacturing strategy (expected maximum temperature and front velocity).