Development of a Draping Algorithm for Non-Structural Aerospace Composites

• Main Author: Hoffer, Jacob
• Other Authors: Robitaille, François
• Format: Others
• Language: en
• Published: Université d'Ottawa / University of Ottawa 2020
• Subjects: Composites; Draping; In-Plane Shearing; Woven Fabric; Simulations
• Online Access: http://hdl.handle.net/10393/40650; http://dx.doi.org/10.20381/ruor-24878

Fibre reinforced polymer matrix composites are used frequently in aerospace applications. Manufacturers of aerospace components favour composites over traditional metallic alloys due to their light weight, high modulus, corrosion resistance and fatigue resistance. Advantages of composites for non-structural interior components over metallic include: ease of manufacturing for single parts of complex geometry as opposed to assemblies, cheaper manufacturing of a limited series of parts and composites greatly reduced noise, vibration and harshness. However, manufacturing interior composite components requires critical attention to detail during the preforming stages and handling of dry fabric textiles. Since these components are handmade they often yield lower profits and therefore efficient preforming is critical. Designing draping strategies for industrial liquid composite moulding processes requires a significant amount of time and testing, in simulation and also working on physical moulds. Mould and part surfaces are often defined by a number of geometric features, labelled base surfaces in the context of this thesis, which can be used to quickly probe multiple draping strategies and identify the best one. Traditionally, trial and error work is performed over a full mould surface until a working or acceptable draping strategy is found, rarely identifying the best strategy. The work in this thesis presents the initial development stages for a draping predictive tool aimed at quickly probing multiple draping scenarios in simulation prior to receiving moulds and identifying the best draping strategy for industrial non-structural aerospace composites. A multi-parameter remodelling tool – the conical frustum – was developed for uniformly identifying base surfaces through 12 geometric parameters linked into a database of in-plane shear and yarn orientations results. The development of the database is discussed, detailing Taguchi methods of experimental design used for developing linear functions from the database results, which allow interpolation of results on base surfaces that do not directly exist within the database. This thesis also includes major developments for the core draping algorithm used for linking individual base surface results together when probing draping strategies. Further investigations were performed on unique elements of in-plane shear behaviour that are encountered during draping, so that these could ultimately be considered during the development of this version of the draping algorithm whilst others may be included in future developments.