Optimisation of low pressure processing for honeycomb sandwich structures

Vacuum-bag processing of honeycomb sandwich structures faces particular challenges compared to autoclave sandwich fabrication for aeronautics, a consequence of using, at the most, atmospheric pressure for their manufacturing. In particular, the adhesion of the skin to the honeycomb core is reduced and the skins tend to have higher values of porosity. On the other hand, autoclave processing is expensive, in particular for large structures, and a growing interest has been observed for vacuum-bag only processing of sandwich structures. During honeycomb sandwich cure, the skin represents a barrier to air flow between the honeycomb and the vacuum-bag so the air transport through the prepreg, adhesive and consumables or other additional materials may be a crucial factor to increase skin debulking, reduce void content and extract air from the honeycomb cells. The pressure inside the honeycomb is, therefore, a result of the through thickness permeability to air of the skins and vacuum pump work. The goal of this thesis is to develop methods for improving the low pressure processing of honeycomb sandwich structures by quantifying the through thickness air permeability of fibre reinforced composites during cure and analysing related processing issues and governing phenomena. An experimental method based on a falling pressure measurement was proposed to evaluate the through thickness air permeability of prepregs during cure, with two set-up variants. One concerns the measurement of the inherent air permeability and the other applies to processing conditions, implying the use of both consumables and vacuum-bag. The methods were successfully tested and fulfil the criterions used in soil science for the evaluation of the air permeability. Prepreg and adhesive permeabilities were determined separately and in skin combination. Prepreg permeability to air during cure varies between 10-19 m2 and 10-17 m2 and was found to depend on the evolution of the resin viscosity, which shifts from an initially immobile phase to a mobile one and then back to an immobile phase. The adhesive film was found to have very low initial permeability, at least two orders of magnitude lower than that of the prepreg. A range of initial skin through thickness air permeability could be achieved by perforating the prepreg plies, the adhesive layer, or a combination of both. An adhesive deposition method which results in the centre of honeycomb cells being free of adhesive was also tested. A corresponding range of achievable pressure was measured inside the honeycomb. The evolution of the through thickness permeability with the curing cycle was determined for each case. The adhesive layer was identified as the element that reduces the initial through thickness air permeability the most in skin manufacturing. This suggests that one of the ways of increasing the initial through thickness air permeability of the skin is to modify the permeability of the adhesive layer. The role of the resulting pressure inside the honeycomb on skin-core adhesion and skin quality was then evaluated. Skin permeability was found to control skin-core adhesion through the pressure drop in the honeycomb cells and potential outgassing of the adhesive layer. An optimal range of pressure inside the honeycomb was found to be between 40 kPa and 70 kPa. A process window is proposed for achieving the optimal pressure inside the honeycomb, as a function of the skin permeability and time of vacuum application. A semipreg alternating dry and resin impregnated areas along the fibre bed surface was also investigated as an alternative to prepreg in skin manufacturing. The semipreg through thickness air permeability before cure is approximately three orders of magnitude higher than that of a unidirectional prepreg impregnated with the same resin, thus allowing the reduction of air pressure in the honeycomb. A model was proposed for the air permeability change during cure, as dry areas get infiltrated. Due to resin pouring inside the honeycomb cells, this type of semipreg is viable as a skin only if combined with a material that has low permeability to resin, e.g., a prepreg. The resulting initial pressure in the honeycomb cells might then be tailored to specific needs. Finally, it was possible to produce sandwich samples combining a consolidation system, such as a prepreg, with an infiltration system, as wet lay-up or with a semipreg. The combination of a prepreg with a semipreg revealed to be a very practical and clean system, where it is possible to have the same resin in both materials. Moreover, it allows to tailor the initial honeycomb pressure as well as during cure, and to have the initial permeability assessed.

Ultra-light photovoltaic composite sandwich structures

Julien Rion

The ultra-light photovoltaic sandwich structure is a new multifunctional structure concept enabling weight and thus energy to be saved in high-tech solutions such as solar cars, solar planes or satellites. The novelty of this approach is to use solar cells as a load carrying element in the structure. The aim of this work was to investigate the failure mechanisms of such ultra-light sandwich structure and their correlation with microstructure, processing pressure, and strength in order to obtain optimal design and processing. To this end, composite sandwich structures were extensively studied with weights in the range of 650 – 850 g/m2, and comprising one 140 µm thick skin made of 0/90° carbon fiber-reinforced plastic (CFRP), one skin made of 130 µm thick mono-crystalline silicon solar cells, and a 29 kg/m3 honeycomb core. As a first step, core-to-skin bonding in a symmetric (CFRP / core / CFRP) sandwich, for which a design criterion was lacking, was especially studied. An adhesive deposition technique was developed enabling the adhesive weight used for core-to-skin bonding to be tailored. Based on adhesive contact angles, the formation of the adhesive fillets between honeycomb cell walls and skin was modeled. Core / skin debonding energy was measured and compared to core tearing energy measured with a new video-based method, and the failure mechanisms during skin peeling were investigated. It was thus ascertained that, to provide the highest debonding energy-to-weight ratio, the optimal adhesive weight was 35-40 g/m2. Furthermore, in contrast with classic sandwich structures with thicker skins, it was observed that the bending strength of the ultra-light sandwich panels increased with adhesive weight. This was due to the formation of adhesive fillets, which significantly increased the bending stiffness of the thin CFRP skin, and thus increased the compressive load causing local instability of the skin. Models taking into account the increased skin stiffness showed that the best adhesive quantity required to increase the strength-to-weight ratio was ∼40 g/m2. In a second step, the influence of processing pressure on the morphology and strength of symmetric (CFRP / core / CFRP) ultralight sandwich structures was investigated by using one-shot vacuum bag processing. This showed that higher processing pressures caused the formation of larger adhesive fillets and an increased waviness of the CFRP skin on vacuum bag side. These two effects had conflicting impacts on the strength of the structure. Waviness of the skin favored local instabilities, whereas adhesive menisci stabilized the skin. Modeling of the local instability of the skin by taking into account the waviness of the skin and the size of the menisci as a function of processing pressure enabled an optimal processing pressure of 0.7 bar to be identified, giving the highest strength-to-weight ratio. The third step of the study was devoted to the mechanical analysis of the mono-crystalline silicon solar cells. The brittle behavior of the cells was confirmed, and the Weibull failure probability curve was established with the mean tensile strength at 221 MPa. It was demonstrated by experimental testing and finite element modeling (FEM) that the low strength of the cells compared to the intrinsic strength of silicon (∼6.9 GPa) was not due to surface texturation of the cells used for increased efficiency, but to more severe surface or edge defects. FEM also showed that no significant reinforcing effect of the cells could be obtained with polymer encapsulation. In addition, thermo-mechanical stresses due to a mismatch of the coefficient of thermal expansion (CTE) between the Si cells and the polymer encapsulation were found to be negligible. In order to protect the cells against the environment, encapsulation of the cells was successfully carried out, using highly transparent fluoropolymer films treated with SiO2 plasma sputtering for improved adhesion, together with silicone adhesive. Finally, the integration of solar cells as a photovoltaic skin of an ultra-light sandwich structure was achieved using thin stress transfer ribbons to ensure load transfer between adjacent cells. It was observed that the cells were not damaged by sandwich panel processing, even in curved panels, thus showing that the processing windows of the different constituents were compatible. The asymmetric (Si / core / CFRP) photovoltaic sandwich structure with a weight equal to 800 g/m2 and a specific power density equal to ∼250 W/kg (i.e. 20 times more than standard commercial photovoltaic panels) demonstrated an equilibrated mechanical behavior, i.e. the CFRP skin, reinforcing ribbons, and solar cells had similar failure loads.

EPFL2008

Film stacking impregnation model for a novel net-shape thermoplastic composite preforming process

Simon Jespersen, Jan-Anders Månson, Véronique Michaud, Martyn Douglas Wakeman

An impregnation model has been developed to evaluate the infiltration phenomena that occur during a novel near net shape preforming method. This process comprises automated deposition of thermoplastic resin and unidirectional (UD) carbon fibres to a pre-programmed stacking sequence, thereby forming tailored preforms for subsequent stamp-forming. Infiltration kinetics have been simulated to study the effect of different stacking scenarios, materials, and preconsolidation routes on the novel preforming process. Isothermal infiltration of a liquid thermoplastic polymer into a compressed UD fibre bed has been examined and the experimental results have been used to validate an infiltration model based on local fluid flow in compressible porous media. This enables simulation of infiltration in alternating matrix film and fibre layers, relating pressure, time, and temperature with the local fibre volume fraction, pressure, and liquid and solid velocities in the stacked material. For a given set of processing conditions, the model fibre volume fraction distribution prediction enables the optimum matrix stacking layer thicknesses to be determined. It was shown that infiltration is inhibited above a limiting pressure which leads to increased fibre bed compaction and hence decreased permeability.

2008

Modeling of axial and shear stresses in multilayer Sandwich beams with stiff core layers

Julia De Castro San Roman Fest, Thomas Keller, Michael Osei-Antwi, Anastasios Vassilopoulos

In order to optimize the load-bearing behavior of sandwich structures and adapt it to specific requirements, their core may be composed of layers of different materials with tailored properties. These layers may consist not only of soft foam or honeycomb materials but also comprise stiff materials, e.g. timber, or even very stiff laminates in the case of fiber-reinforced polymer (FRP) materials. In this paper, new analytical models for predicting axial and shear stresses in multilayer sandwich structures composed of stiff core layers and intermediate laminates are presented. The models are based on new formulations for calculating the bending and shear stiffness of multilayer sandwiches. They have been validated by finite element modeling (FEM) and the results from four-point bending experiments on GFRP-balsa sandwich beams with complex core assembly. In contrast to existing models, e.g. the high-order sandwich panel theory (HSAPT), the new models are able to accurately predict axial and shear stresses in stiff cores and intermediate FRP laminate layers. (C) 2014 Elsevier Ltd. All rights reserved.

Elsevier2014

Ultra-light photovoltaic composite sandwich structures

Julien Rion

EPFL2008