Improving the process modelling capability for manufacturing large composite structures used on passenger aircraft
thesisposted on 01.03.2017, 04:12 by Pierce, Robert Samuel
As the aircraft manufacturing industry continues to transition from metals to composite materials for structural applications, it is becoming increasingly difficult and costly to manufacture components with traditional autoclave processes. The demand for large, complex and low-cost composite aerostructures is prompting advances in manufacturing with textile reinforcement materials and Liquid Composite Moulding (LCM) techniques. However, these methods continue to rely on operator skill and experience through empirical practices. In order to reduce production timescales and resources, this research focuses on the development of a Complete Process Model (CPM) that can simulate LCM, and emphasises the impact of dry fabric deformation on subsequent infusion. First, the fabric deformation that results from draping was simulated using a hypoelastic continuum approach in the finite element package, Abaqus/Explicit. The strength of this model, which has been validated against existing simulated and experimental results, is attributed to the non-orthogonal tracking of fibre directions through the use of a user subroutine (VUMAT). The tensile and shear properties of a carbon fibre, plain weave material were characterised experimentally in order to enhance the fidelity of fabric behaviour in the draping model. For improved shear measurements, an in-house Digital Image Correlation code was developed for use with a digital camera to monitor fabric deformation. Radial permeability characterisation testing was also performed over a range of shear angles to capture the complete, deformation-dependent, anisotropic permeability behaviour of the aerospace-grade fabric. Video footage of the fluid flow was digitally captured and processed, using a statistical approach, in order to define permeability as a function of both shear angle and flow orientation. The infusion stage of the simulation employed a Volume of Fluid method in ANSYS FLUENT to model transient, multiphase flow through an anisotropic porous medium. The approach was validated against experimental test results. These components were combined to form the Complete Process Model. The material characterisation provided the necessary data for the draping model to produce a realistic prediction of the deformed fabric geometry and shear angle distribution. This information was then coupled with the results of the permeability characterisation experiments to create a complex permeability distribution, before being passed on to the infusion model. The CPM demonstrates the importance of linking draping deformation with infusion, shows significant improvement over traditional modelling, and serves as a solid foundation for further advances.