Monday, May 20, 2013
    
Multi-scale modelling to predict defect formation during resin infusion

Principal Investigator

A Long (Nottingham)

Co-investigators

R Brooks (Nottingham), M Wisnom (Bristol), I Partridge (Bristol)

Names of academic partners (with cash/in-kind contribution)

University of Nottingham (lead), University of Bristol (studentship), Cranfield University (studentship)

Names of any industrial partners (with cash/in-kind contribution)

Rolls-Royce, ESI, LMAT, Vestas, Airbus

Chopped carbon fibre preforms for LCM processing of automotive components, generated automatically by robot, show random distribution of fibres

Summary of project

This project aims to develop and validate a multi-scale modelling approach to predict variability and defect formation during liquid moulding, capturing the combined effects of material, geometry and process variables. This will incorporate variations in materials and processing behaviour at the macro-, meso- and micro-scales.

Objectives of project in order of priority

  • To develop meso-scale, steady state simulations to determine base reinforcement permeability as a function of fibre volume fraction and stochastic variation in tow paths.
  • To develop meso-scale, transient simulations for flow and cure of a reactive resin system to determine the range of defects that can occur as a function of fibre volume fraction and statistical variation.
  • To develop macro-scale, transient simulations for a reactive resin system utilising predicted permeability distributions, allowing the range of possible flow outcomes to be predicted.
  • To develop structural FE analysis for meso-scale models containing statistically generated meso-and micro-defects, predicting effects on mechanical behaviour (static, fatigue) utilising a damage mechanics approach.
  • To validate predicted defects at all scales and their effects on mechanical performance. CT scanning and microscopy will be used to establish the detailed morphology of real voidage defects, with the impact on mechanical performance under different loading modes established from coupon testing. This will provide validation for both the defect modelling and FE analyses, along with a broader understanding of the morphology of real defects.
Progress to Date Minimize
In the first year we have made significant progress in understanding the micro-scale behaviour of reinforcements during manufacture by resin infusion. We have characterised the fibre distributions within tows (fire bundles) within composite reinforcements, and have developed a detailed statistical description of these. This has been used to generate virtual models, which have been used within CFD studies to simulate flow during infusion. Steady-state simulations allow the distribution of tow permeabilities to be predicted, and here a key discovery is that the mean value is almost an order of magnitude lower than that predicted by (previously accepted) analytical models. Transient flow simulations are now starting to determine the conditions for void formation.
At the meso-scale (unit cell level) we have used MRI to image the flow of fluid through reinforcement during infusion. This is the first application of this approach to composites manufacture, and we have used this to identify void formation between tows.
Work on cure modelling will also focus on variability, and here a stochastic simulation methodology is being developed to study the uncertainties/ performance tradeoffs in cure processes. Initial work has focused on reviewing the key sources of variability and implementing a robust methodology to model composite cure and distortion that allows efficient incorporation of variability information. The work currently focuses on the development of a Monte Carlo scheme for uncertainty simulation of cure and the development of a sensing setup that will allow the estimation of surface heat transfer variability in an industrial environment. 

Downloads
Copyright 2013 EPSRC   |  Built by NetComposites Sitemap  |  Privacy Statement  |  Terms Of Use