Modelling of the CVI Processes

Abstract

The chemical vapour infiltration (CVI) process is used to fabricate the interphases and matrices of CMCs. This process involves complex physico-chemical phenomena such as the transport of precursor, carrier, and by-product gases in the reactor and inside a fibrous preform, chemical reactions (pyrolysis and deposition), and the structural evolution of the preform. It is able to provide high-quality materials because the process conditions are rather mild with respect to the fibres ; however it is expensive and sometimes difficult to optimize. Many variations of the basic concept have been proposed in the past decades, introducing thermal and pressure gradients, in order to increase the efficiency. This process has been the object of extensive modelling efforts, because of imperative optimization needs. The present work is an attempt to provide a synthetic view of these models, focusing on several features of CVI modelling : i) Modelling CVI requires a multi-scale strategy, with models ranging from process scale down to atomic scale, ii) Original physicochemical couplings are involved, which require the development of adequate treatments, iii) There is a hierarchy of model refinement, ranging from fully detailed models to quasi-analytical predictions.

Fibre-scale Modeling of C/C Processing by Chemical Vapour Infiltration using Xray CMT Images and Random Walkers

“…Application of this procedure to various preforms allows discussing their relative merits in terms of infiltrability, i.e., their ability to receive a matrix deposit in given conditions. Moreover, the simulated evolution results may be plugged inside a more global CVI modeling strategy, including a more detailed chemistry and a reactor-scale description of the mass and heat transfer (35). For instance, by combining the results of this approach with a chemical deposition kinetic scheme identified from CVD experiments inside a reactor-scale Finite Element model, it has been possible to get a quantitative validation of the whole modeling procedure in the case of carbon foam reinforcement (36).…”

Image-Based Numerical Modeling of Self-Healing in a Ceramic-Matrix Minicomposite

“…This process gives characteristic continuous layers growing in conformal shape with respect to the initial substrate (i.e., the fibers). The layer thicknesses are larger in the outer parts of the fiber bundles, because of the competition between chemical deposition reactions and gas diffusion [17]. As a result, some pores remain between the fibers after infiltration.…”

“…The solution of the fluid spreading equations (Equations (16) and (17) or (17)) was achieved with an inexpensive approximate numerical scheme, consisting of distributing uniformly the amount of fluid created in any connected component of the liquid phase (a "droplet") during a time increment ∆t on all neighboring nodes surrounding the existing droplet. The continuity Equation (17) in its integral form can be written as a mass balance over ∆t:…”

Rarefied Pure Gas Transport in Non-isothermal Porous Media: Validation and Tests of the Model

“…Low pressures are also common in CVI, since it is a convenient way to reduce the Thiele modulus ; accordingly, the Knudsen number is usually not negligible. A large amount of modelling efforts have been made (Starr and Smith, 1990;Ofori and Sotirchos, 1996a;Reuge and Vignoles, 2005;Vignoles, 2006), featuring or not Knudsen transport, but few has been said about thermal transpiration in these works.…”