Prediction and characterization of the resin flow and the fibrous preform behaviour during Liquid Resin Infusion (LRI) S. Drapier 1 , J. Molimard 1 , P. Wang 1 , A. Vautrin 1 , P. Henrat 2 1 LTDS / UMR 5513 / ECL / ENISE / ENSM-SE SMS, ENSM-SE, 158, cours Fauriel - 42023 Saint-Étienne cedex 2 – France 2 HEXCEL Reinforcements, ZI les Nappes 38630 Les Avenières KEYWORDS: liquid composite moulding, liquid resin infusion, resin flow, polymer matrix composites INTRODUCTION Weight saving is still a key issue for aerospace industry, therefore CFRP prepregs have been extensively utilized in the B787 and A350 programmes to lighten the structures. 50% of those aircraft structures is made of high performance composite materials. To reach a mass ratio of 60%, it is now necessary to manufacture thicker parts which are submitted to severe mechanical service loading. Prepregs are no longer suited to that type of component parts and advanced manufacturing techniques like resin injection (RTM) or infusion (LRI, RFI) of fibrous preforms should be used. However, since those parts are security parts and have complex shapes, numerical simulations of the process should be performed to optimise the design and manufacturing parameters and reduce the defects. A complete model of infusion-based processes has been recently developed by Celle et al. [1]. It gives the opportunity to develop simulation of real infusion process, taking account of large variations of the preform thickness during the process. The model has shown its ability to deal with the highly non-linear mechanism of resin infusion into fibrous preforms which undergo high compaction. Obviously, there is a need to set up a reference and representative experiment to check the performances of the model in 2D and identify the material parameters that are required to carry out reliable simulations. Major points to check first is the resin flow through the preform and the thickness variation during resin infusion. A special experiment has been build up to manufacture plates of different thickness by LRI. EXPERIMENTAL APPROACH AND ANALYSIS The numerical model makes it possible to predict the resin flow during the infusion process by postulating that the resin first fills in the drainer on the upper surface of the preform, and then infuses gradually in the thickness direction. The fluid front detection will allow both to verify the two-stages flow: in-plane flow through the drainer and flow through the thickness, and correlate the numerical flow front position. Since the temperature of the liquid resin is significantly lower than that of the preform, it is shown that local temperature measurements by using micro-thermocouples takes full advantage of this point and can lead to a very first validation of resin flow assumption through the thickness of a plate preform. Prediction of the thickness of the preform is also carried out, however the only information that can be extracted from experiments at the moment are the initial and final part thicknesses. The final thickness results both from the infusion process and from the resin curing that is known to induce a volume shrinkage. In order to decouple the infusion stage from the curing step and take advantage of the amount of information available from the simulation, a continuous thickness measurement is proposed. This measurement is based on fringe projections on the preform stacking undergoing infusion [2]. The deformed surface of the bag