Virtual un-manufacturing of fibre-steered preforms for complex geometry composites

Home > Research > Completed

Virtual un-manufacturing of fibre-steered preforms for complex geometry composites

Host Institution: The University of Bristol

Start Date: 1st April, 2019

Duration: 6 months

Lead Investigator: Stephen Hallett

Co-Investigators: Eric Kim, Jonathan Belnoue

Aims

In most cases, complex geometry composites components are designed based on ideal or theoretical fibre angles, with little or no consideration of the manufacturing processes or constraints involved in delivering them. This feasibility study aims to take a novel virtual approach to “un-manufacture” these ideal designs for the case of formed composites, so that a flat, tailored preform can be created, which results in the required ideal fibre architectures after forming (see Fig. 1). The primary manufacturing process envisaged to deliver this is diaphragm forming of thermoset prepregs, deposited using automated deposition process. However, it is anticipated that the concept developed will be applicable to textile preforms (including non-crimp fabrics) and thermoplastic prepregs.

This novel approach of manufacturing a flat fibre-steered preform and forming afterwards can offer significant benefit in terms of production speed for complex composite components. The equipment involved can be less complex and therefore also cheaper and more generic. The key enabler is a virtual simulation capability to accurately predict the required fibre paths of the un-manufactured preform. This proposal therefore feeds into both the Hub grand challenges on process robustness enhancement through a better understanding of the underpinning science and the development of high-rate processing technologies. It also delivers on two of the Hub priority areas, namely rapid processing technologies and design for manufacture via validated simulations.

Delivering this vision through validated simulation will help speed up the process development in comparison to more traditional trial-and-error methods based on costly and time-consuming experiments. It will help minimise the effort associated with rework of the process such as changes of geometry or materials.

Figure 1: Schematic of the lay-flat and form.

Novelty Automated fibre placement (AFP) has led to considerable efficiency gains in the manufacturing of relatively flat mechanical parts made of composites. However, fast deposition of material on complex doubly-curved surfaces remains very challenging. In fact, the required fibre steering in order to conform the material to the complex mould shape can only be performed at a very low speed and often results in defects such as fibre buckling under the effect of in-plane tape bending in the case of low steering radii. The maximum linear deposition speed of the AFP machine becomes no longer a metric of production rate.

Figure 2. Comparison between experimental results (left) and model prediction (right) for the forming a multi-layer dry-fabric on
a tetrahedral tool.

Figure 3. Wide tape CTS machine steering a 100 mm wide unidirectional prepreg tape.

Another automated manufacturing process widely used in the industry, is the diaphragm forming of continuous fibre prepregs. This allows fast and cheap production of relatively simple composite parts. However, producing defect-free parts with high level of double-curvature is extremely difficult, as the combined effect of compressive loads in the fibre directions and stiffening of the material at high shear angles, which develop when the stacks of anisotropic prepreg sheets are forced to adopt the shape of the tool, often result in the formation of out-of-plane wrinkles and fibre path deviation.

A novel manufacturing approach proposed here is a hybrid method to combine the benefits of both automated fibre deposition and diaphragm forming. The key focus of the feasibility study will be to develop a numerical simulation method to reverse-engineer the fibre paths of a target draped shape into a flat undraped fibre-steered preform. Starting from a desired fibre path on a complex shape, flat preforms will be extracted through an un-manufacturing simulation to be developed in this project. This will build on the work performed under two successive EPSRC projects (i.e. CIMComp core project – DefGen, and the platform grant SIMPROCS) in which Prof. Hallett (SH) and Dr Belnoue (JB) have co-developed world-leading validated predictive capabilities of the mechanical response of prepregs under processing conditions and forming-induced wrinkles in stacks of fibrous reinforcement, as illustrated in Fig. 2. The designed flat preforms will be manufactured using a cutting-edge fibre steering technology developed by Dr Kim (EK). This advanced tape laying process is a world first Automated Tape Laying (ATL) process with defectfree fibre steering capability using the Continuous Tow Shearing (CTS) mechanism, which was developed during his fellowship funded by the predecessor of Hub (i.e. the EPSRC CIMComp grant). As illustrated in Fig. 3, CTS allows for steering fibres within a 100 mm wide unidirectional carbon/epoxy prepreg tape without causing defects by continuously shearing the tape. Its minimum steering radius is about 50 mm, which is an order and several orders of magnitude less than those of modern AFP and ATL machines, respectively. The obtained preforms will then be diaphragm formed at elevated temperature to mitigate against wrinkle formation by exploiting the reduction of prepreg shear-stiffness observed when the resin viscosity is reduced.

Copyright © 2020 CIMComp