Street, George Edward
(2025)
Thermomechanical forming simulation for fibre reinforced thermoplastic laminates.
PhD thesis, University of Nottingham.
Abstract
The use of fibre reinforced composite (FRC) materials has grown exponentially in recent years, due mainly to the attractive properties that they exhibit compared with conventional isotropic materials. While they pose an attractive proposition, challenges to the composite industry stem from long manufacturing time, poor recyclability and difficult joinability. Fibre reinforced thermoplastic (FRTP) composites aim to meet the challenges posed by thermoset FRCs. The increase in demand for FRTP components has driven the need for improved continuous sheet forming processes allowing for viable production of high-performance components at a range of manufacturing scales. This includes improved process modelling, with the objective to reduce the reliance on costly and time-consuming trial-and-error methods and optimise process control.
The aim of this work was to produce a functional FRTP thermomechanical forming simulation for the attractive double-diaphragm forming (DDF) process, identified as a gap in literature where current FRTP models are limited to just the matched-tool processes. It was envisaged that the model would allow for accurate defect prediction at a range of forming conditions, subsequently allowing for optimisation of the FRTP DDF process. To achieve this aim, the thesis was broken down into four key areas:
(i) FRTP DDF Experimental Study
A FRTP DDF rig was produced to conduct forming at a range of conditions. The objective was to identify the parameters for successful forming, such as materials and forming conditions, and also provide validation for the subsequent numerical model. Successful isothermal forming was conducted with single-ply carbon fibre (CF) / Polyamide 6 (PA6) organosheets, utilising silicone diaphragms over a double-dome tool geometry. For this purpose, preheating of the diaphragms to 200 °C was necessary to mitigate issues resulting from the significant thermal expansion of silicone. It was shown that both reducing the laminate temperature and/or increasing the forming rate results in increased defect formation in the formed component. For example, bridging (tool-laminate conformity) and wrinkle lengths increased by 110% and 490% respectively with forming at 200 °C compared to that at 250 °C. High-rate forming resulted in a 58% increase in wrinkle amplitude. It was found that using an infrared (IR) lamp resulted in significant in-plane thermal distributions of up to 18 °C, and that the solid organosheet did not permit airflow, thus inhibiting vacuum propagation at the start of the forming process. A maximum diaphragm strain of 80% was noted, highlighting that both laminate and diaphragm deformation are critical to the success of the DDF process.
Experimental repeatability was analysed and found to be high at higher temperatures, with variations in bridging and shear generally limited to 8%. This repeatability degraded at lower temperatures, especially concerning wrinkle propagation, both between repeats and intra-test between two sides of a symmetrical tool. This highlighted the instability of wrinkles and the variability that can occur owing to errors with specimen alignment and heating.
(ii) Material Behaviour Characterisation
It was necessary to characterise the significant deformation mechanisms, associated with both the laminate and diaphragms, which allowed mould conformity to be achieved in the experimental DDF routine. The significant laminate deformation mechanisms that were analysed were deconsolidation, intra-ply shear, out-of-plane bending (intra-ply mechanisms) and finally the interaction between the laminate and the silicone diaphragms. Deconsolidation was found to be limited to 9% upon heating, owing to the single-ply laminates utilised in this work. A critical reconsolidation pressure of 0.4 bar was also identified, therefore indicating that full reconsolidation is achieved in the DDF process. Intra-ply shear characterisation was conducted predominately utilising the bias extension test due to the poor repeatability (45% variations) of the picture frame test. This was caused by poor specimen alignment (non-orthogonality in the organosheets) and localised effects from the clamping arrangement. Shear forces were shown to increase by approximately 440% and 240% with a temperature reduction from 270 °C to 210 °C and 25x forming-rate increase, respectively. This behaviour was parameterised using an overstress-law and an Arrhenius-type relation respectively. A similar parameterisation process was utilised for out-of-plane bending, with experimental tests conducted with the cantilever bending test. It was found, however, that the influence of bending-rate could not be analysed with this technique. The diaphragm-laminate friction was characterised with the pull-through method, and parameterised with a Stribeck analysis, showing the friction condition to lie within the hydrodynamic regime. Regarding the diaphragms, hyperelastic behaviour was characterised with the uniaxial, biaxial and pure-shear methods. Silicone was found to not exhibit significant directionality or rate-dependence (within 6%), but did exhibit temperature-dependence and was influenced by the loading history. To remove the influence of loading history, a novel technique was utilised where the samples were pre-compressed before testing. This also resulted in an approximately 25% reduction in material stiffness.
(iii) FRTP DDF Thermomechanical Simulation
Following the successful characterisation of the material behaviour, it was implemented within a DDF numerical routine. The model itself was based on a continuous hypoelastic formulation with decoupled membrane and bending behaviour. Initially, each deformation mechanism was validated using Abaqus/Explicit to ensure that it was captured accurately. Intra-ply shear viscoelasticity was well captured, although a gradually increasing error of 20% - 100% was present due to the normalisation procedure adopted. Diaphragm-ply friction behaviour was captured within 16% of the experimental pull-through forces and included a novel method for incorporating the initial ‘static’ friction within the model. The validity of the bending input could not be validated due to buckling that was identified on the experimental samples, however a sensitivity study showed it was captured sufficiently within defined bounds for accurate wrinkle prediction. Finally, the Ogden hyperelastic material model for the silicone diaphragm was shown to accurately capture the elastomeric behaviour within 10% of the experimental forces. The isothermal DDF model was subsequently produced to replicate the experimental forming process, run at the different process conditions adopted and then validated against did, nonetheless, predict the experimental forms. At higher temperatures (210 °C ≤), the model was able to capture bridging within 0.5 mm, wrinkle amplitudes within 20% and maximum shear angles within 1°. The accuracy of the model was also shown to not degrade with an alternative -45/45° fibre orientation. At a higher forming rate, wrinkling was still well predicted, although a slight (0.9 mm) overestimation of bridging was identified. The model nonetheless predicted the occurrence of secondary wrinkling in the high shear regions at this rate, and also predicted this behaviour well for the lowest tested temperature of 200 °C. Bridging was significantly underestimated at this temperature, however, caused by the inhomogeneous in-plane thermal profile in the experimental tests. This profile was subsequently incorporated, resulting in a 44% and 62% improvement in bridging and shear prediction respectively, at lower temperatures.
(iv) FRTP DDF Optimisation
Having produced a functional FRTP DDF simulation that could predict forming behaviour with a good degree of accuracy, the model was subsequently adapted to allow for optimisation of the process conditions. Initially, the heat transfer mechanisms of conduction, radiation and convection were obtained through numerical, experimental and analytical methods respectively. Each of these methods was subsequently validated within an Abaqus/Explicit numerical model, with laminate cooling captured within 3% accuracy. This was coupled in parallel with modelling of laminate crystallinity, identified in the initial numerical model as an important consideration for low-temperature forming. This was twinned with crystallisation-dependent specific heat capacity, such that the crystallisation exotherm was captured in the model. The influence of crystallisation was implemented utilising a rule of mixtures approach between the molten and solid material states. Non-isothermal simulations were run at different forming rates, and it was found that including transient effects results in a slight bridging reduction (0.3 mm) at higher forming rates, as opposed to the increase (0.1 mm) noted in the isothermal study. Wrinkle amplitudes were also more aligned (within 4%) between the two non-isothermal forming rates, compared to the isothermal equivalent (72% differential). This was followed by a tool temperature optimisation process where the tool temperature was minimised for the lowest demoulding time, yet still forming within acceptable defect limits. A minimum tool temperature of 200 °C was identified, resulting in bridging of less than 5 mm and reduced wrinkling behaviour, with a demoulding time of 38 seconds. Finally, a novel ‘zoned’ tool heating strategy was implemented, allowing for equivalent forming quality but resulting in a 42% reduction in the demoulding time.
In summary, this work provides the foundations for accurate FRTP forming simulation of alternative manufacturing techniques to conventional matched-tool methods, with the possibility for increased process monitoring, control and optimisation.
| Item Type: |
Thesis (University of Nottingham only)
(PhD)
|
| Supervisors: |
Johnson, Michael Sylvester
Ou, Hengan |
| Keywords: |
Fibre reinforced thermoplastic composites; Thermomechanical forming; Process optimisation |
| Subjects: |
T Technology > TS Manufactures |
| Faculties/Schools: |
UK Campuses > Faculty of Engineering > Department of Mechanical, Materials and Manufacturing Engineering |
| Item ID: |
81990 |
| Depositing User: |
Street, George
|
| Date Deposited: |
31 Dec 2025 04:40 |
| Last Modified: |
31 Dec 2025 04:40 |
| URI: |
https://eprints.nottingham.ac.uk/id/eprint/81990 |
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
