| Geometric modeling of woven preforms is a useful tool to predict preform thickness, preform areal density and fiber volume fraction (FVF) of constituent yarns. Previous geometrical models of 3D orthogonal woven preforms, which are extensively reviewed in Chapter 2, were limited to plain weave interlacing pattern in jammed case. In this study, generalized geometric models in terms of weave design (represented by a numerical value termed "weave factor") were developed. The models cover both jammed and non-jammed cases, consider circular, racetrack, and rectangular yarn cross-sectional shapes. The models predict thickness, constituent yarn weights, and FVFs of 3D orthogonal woven preforms. The models illustrated fabric architecture potential of 3D orthogonal woven preforms. Numerical results for hypothetical structures showed how to control through the thickness components of the z-yarn and total FVF, that have direct effect on the in-plane and out-of-plane properties, with interlacing pattern (weave factor) and z-yarn linear density. The models were demonstrated as an essential design tool that may be used to develop composites with predicted level of structural parameters and performance.;Broad range of 3D orthogonal woven preforms from glass fibers with different architectures were woven and consolidated by vacuum infusion process (VIP) with different z-yarn interlacing pattern, number of y-yarn layers, and x-yarn spacing to verify the model for filament yarns.;Dry preform thickness and weight of in-plane yarns predicted by the geometric model for filament yarns correlated well with experimental results. Z-yarn weight of dry preform was 24.3% overestimated by the model due to shortening of z-yarn at cross overs in real preforms due to the flattening of x-yarns caused by the tension of z-yarns. Total FVF of actual dry preform was 0.4% greater than model prediction. However, total FVF of composite was 5.4% overestimated by the model, which is within the experimental error.;In VIP, presence of flexible vacuum bag on one side of the molding and pressure gradient from resin inlet to flow front change the compaction pressure on unsaturated preform that results in thickness and FVF variation in cured composite. Previous work on composite using VIP dealt with resin flow modeling to predict thickness, FVF, and resin pressure profiles. However, the effect of these profiles on mechanical properties was not investigated. In this study, specimens from composite panels were tracked in terms of their positions relative to resin inlet. Analytical (thickness, FVF, and void content) and mechanical (tensile, three-point flexure, and impact) properties of the specimens were measured to investigate the effect of specimen location on composite dimension and performance. Additionally, vacuum (75 and 100 kPa levels) was considered as variable to investigate its effect on composite formation and performance. Effect of fabric architecture on resin infusion performance was demonstrated by faster mold filling times of preforms with plain and basket weaves due to their straight, uniform, and wide resin flow channels compared to tortuous channels in preforms with twill weaves. Analytical properties of 3D orthogonal woven glass fiber composites showed parallel result with previous VIP researches such as thickness decreased, FVF increased and void content fluctuated from resin inlet to outlet.;Slight increase from resin inlet to outlet was observed for the tensile properties in x- and y-directions. Due to their higher FVF compared to the composites with other two weaves, composites with twill weaves resulted in better mechanical properties. Increase in x-yarn density caused increase in tensile stress in x-direction, whereas it resulted in reduction in tensile properties in y-direction. Increase in vacuum pressure slightly improved tensile stress in both directions.;Peak tensile stress in y-direction of three layers, 5.48 x-yarns/layer/cm balanced (defined as identical total denier in x- and y-directions/unit sample width) composites manufactured at 100 kPa vacuums was 15.24% greater than peak tensile stress in x-direction due to more uniform placement of y-yarn than x-yarn. (Abstract shortened by UMI.). |
