This paper was produced for the 2019 NAFEMS World Congress in Quebec Canada
Braiding is a commonly used process for continuous production of composite parts. Relative to similar preforming techniques, braiding has a high deposition rate and high level of automation, so is appealing for high volume production of components between 1m and 10m in scale. To reduce manufacturing process design costs, several braiding simulation tools have been developed in the last decade, with kinematic and finite element models providing the most common simulation strategies.
Kinematic modelling is a fast, cost-effective simulation approach. It is purely based on the geometry of the mandrel and the motion of the tows. In its basic form, it does not account for forces within the system: for example friction, tension, or weight acting on the tows. These parameters have a strong influence on the process, and a finite element model that considers them would be more capable of quantifying their impact on the microstructure of the braid, for example the braid angle and bridging behaviour.
To the best knowledge of the authors, the availability of commercial software for braiding simulation remains limited, hindering the wider use of process simulation in industrial practice.
The National Composites Centre (NCC) has acquired a 288/192-carrier overbraider for the manufacture of large-scale components, up to 800mm effective diameter. In order to increase manufactured output quality whilst reducing the operational costs of physical trials, the NCC has developed an integrated simulation approach that combines a kinematic braiding tool, available on the market, and an in-house finite element model of the braiding process, developed in collaboration with the University of Bristol.
The kinematic tool is Composite Braiding Modeler (CBX), available as a dedicated workbench within Catia platforms. The finite element model is a parameterised Abaqus model, which can be fully customised by the user to replicate the machine configuration.
The approach herein proposed starts from the kinematic simulation of braiding. Given pre-selected process parameters, CBX is capable of optimising the mandrel velocity profile to maintain a desired braid angle. This mandrel velocity profile is then used as input for the finite element model. The output of the two models are compared. In particular, the difference between designed and predicted braid angle, and the prediction of bridging behaviour is used to measure the effect of friction and tow tension on the final predicted braid architecture.
In this paper, this approach is applied to a large axisymmetric component with varying cross-section. It will be shown how the two approaches combined can identify manufacturing defects and provide mitigation actions, with a substantial reduction in process costs and waste.