| Vacuum bag-only processes, used in the manufacture of composite structures (fiber reinforced polymer), are inexpensive alternatives to similar high pressure processes which can reliably produce highly consolidated---high fiber volume fraction, low void content---composite laminates. Prepregs---fiber plies, pre-impregnated with resin (thermoset or thermoplastic)---are processed with elevated temperature and pressure (applied normal to the laminate), which consolidates and cures the prepregs to form the composite structure. Research within the last decade has proven that low void content composite laminates can be produced using prepregs with a vacuum bag only process (in ambient atmosphere) if the prepregs are only partially impregnated with resin.;Open porosity (non-impregnated cross-sections) of the fiber architecture, serves as air evacuation pathways, which allows vacuum pressure applied at the boundaries of the laminated structure to evacuate any gases before becoming entrapped in the resin. The nature of resin distribution (and redistribution), evolution of open porosity, and it's effect on gas evacuation---which intrinsically defines the material properties pertaining to processing and process outcome---is poorly understood. This dissertation pursues a framework and methodology to characterize the relationship between resin saturation and macroscopic gas evacuation properties (permeability, porosity, and Klinkenberg effects), as well as the near-microscopic flow of resin during consolidation.;This dissertation introduces an in-situ resin visualization method used to (i) model the dual scale resin impregnation as a function of pressure and temperature, (ii) observe and model the movement of bubbles which travel with the resin toward evacuated air pathways,and (iii) quantify the surface saturation, which is used to characterize gas permeability as it changes with resin saturation.;This surface visualization method demonstrates that resin flow observed to strongly follow a dual scale flow pattern. A flow model is introduced to describe the two observed flow stages: inter-fiber tow flow and intra-tow flow. By matching the experimental data with the model, values of permeability are estimated from inter-tow pores and intra-tow pores.;Using the same visualization technique, the focus is changed to track bubbles flowing in the resin. Bubbles are observed to emerge through pinholes and flow with the resin through inter-tow channels. A key finding of this study is that tunable process parameters, such as pressure and temperature, are less important for successful bubble removal as compared to the initial state of resin impregnation in the prepreg. Prepregs with high resin impregnation will not be able to vent bubbles, but with sufficiently low resin impregnation, bubbles may escape into air pathways. Small Capillary number theory (i.e. Ca < 0.01) was shown to under predict the relative velocity of bubbles, suggesting that surface tension does not significantly contribute to the drag force on bubbles.;Gas evacuation from a partially impregnated prepreg, was characterized using the pulse-decay method. Dimensionless analysis was used to show how the initial pressure, boundary conditions (vacuum pressure at x = 0 and with or without a reservoir volume at x = L), and Klinkenberg parameter effect the predicted decay of gas pressure. A universal scaling function was identified, which predicts the decay of gas pressure from an empty reservoir volume, through the porous material. By comparing the experimental data to a set of dimensionless master curves, the intrinsic permeability, Klinkenberg parameter, and porosity were determined. |
