Characterisation and statistical regeneration of 3D printed continuous carbon fibre composite structures

Abstract

The process of layer-by-layer constructing a product is known as additive manufacturing. Manufacturing Continuous fibre reinforced composite through Additive manufacturing (AM) technique provides a lot of benefits. One major benefit is its capability to produce low-cost composite structures by optimisation i.e., placing fibres only in required region while manufacturing a layer. However, before optimisation, the AM composite must be predicted accurately through Finite Element Analysis (FEA). Currently, FEA is unable to predict accurately because of limited knowledge in microstructural properties of AM part. Here we propose a methodology which bridges the gap between microstructural properties and FEA model.

The proposed methodology in this study has three major steps i.e., Printing, image acquisition/ processing and regeneration. The proposed methodology was developed and tested on PLA print paths and the same was then repeated for Continuous Carbon Fibre (CCF) paths. The first step includes printing the paths on a detachable print bed. The print bed is then taken to the digital microscope for image acquisition. The microscopic image is further processed to determine deviation in centroid and width of the print paths. Deviation is characterised based on systematic and stochastic type. On the basis of characterisation results, the centroid and width of the paths are statistically regenerated using Markov Chain Monte Carlo (MCMC) method. Since 45°, 90°, 135° and curve paths are commonly observed in AM part, paths with these geometric features are characterised and regenerated using the proposed methodology. Also, this study discuses on the printing of CCF using Fused Filament Fabrication (FFF) i.e., modification done in the 3D printer to achieve CCF printing.

The microstructural study revealed that the deviation in print paths differ with different geometric shapes, ultimately affecting the overall strength. The regeneration results from PLA and CCF paths shows that this methodology was able to regenerate paths with 93-98% of resemblance with the actual paths. The statistically regenerated paths could further help to build a virtual model, that can be used in predicting any AM part’s performance and further optimise the same to lower its cost.

Date of Award	28 Mar 2024
Original language	English
Awarding Institution	Northumbria University
Supervisor	Matthew Blacklock (Supervisor) & Martin Birkett (Supervisor)