Experimental characterisation and forming analysis of pure unidirectional stitched glass fibre non-crimp fabric

Kahavitage Dona, Himanthi Nimrekha Kahavita (2025) Experimental characterisation and forming analysis of pure unidirectional stitched glass fibre non-crimp fabric. PhD thesis, University of Glasgow.

Abstract

This research focuses on the experimental characterisation and forming analysis of pureunidirectional non-crimp fabric (pure-UDNCF). This fabric consists of low-stiffness polyamide stitches with a tricot-chain stitching pattern where the chain stitches run along the front of the fabric, perpendicular to the glass tow direction. Notably, there are no stabilising tows oriented transverse to the main tow direction in this fabric, a common feature in many ‘quasi-UDNCF’, this allows extension of the stitch in the transverse direction under certain loading conditions. The lack of stabilising tows introduces a possible low-energy deformation mode to the pure-UDNCF, which is absent in biaxial fabrics and primarily in quasi-UDNCF. Thus, compared to biaxial fabrics, the inplane deformation modes of pure-UDNCF are complex and dominated by transverse extension and shear. Therefore, characterisation tests typically developed to characterise woven fabrics are unsuitable for pure-UDNCF, because more than one low-energy deformation may occur, and the deformations can be coupled during deformation experiments. This was evident following the initial evaluation of pure-UDNCF with two principal shear tests, namely the Uniaxial Bias Extension test (UBE test is a tensile test performed on a rectangular-shaped fabric with the warp and weft rows oriented ±45° to the direction of applied tensile force) and the Picture Frame (PF) test (all four sides are clamped orthogonal to the tows). The results revealed a significant difference in shear stiffness, with the PF test being significantly higher than the UBE test. This unique observation encourages further investigation to determine the cause of the difference. The intent is to understand whether the different behaviour of the two tests is real, or a problem related to using the tests for this fabric.

The experimental error of the PF test was investigated using two modifications: the pre-displaced PF test and the G-clamped PF test. The pre-displaced PF test was used to investigate the misalignment error and involved two pre-displacements (4mm and 6mm). The existing tight clamps (nut and bolts) were replaced with G-clamps of two different pressures (high and low) to determine how the clamping condition affected the measured shear force. The resulting combined (high and low) G-clamped PF test curve was close to the 6mm pre-displaced PF test curve, and both curves were significantly reduced than the initial PF test results. Furthermore, the friction in the bearing of the PF rig was eliminated to improve the accuracy of the PF test, however, even after these modifications, the standard UBE test results showed less axial force than the PF test curve. Moreover, in addition to shear, an in-plane bending contribution was identified at high shear angles in the PF test. As a result, the axial force measured in the PF test of the pure-UDNCFs was considered to be a combination of shear and in-plane bending. In the UBE test, the in-plane contribution was not significant, however, stretching in the stitch direction was observed and assumed to help minimise the contribution from in-plane bending of the tows. Therefore, during the UBE test, was considered to be a coupling between shear stiffness-tensile strain in the stitch direction, and between tensile stiffness in the stitch direction-shear strain. Experiments were further performed to isolate the contribution of each deformation mode i.e., shear, in-plane bending and tensile strain in the stitch direction. New characterisation tests namely cruciform bias extension (CBE) test, parallelogram shear-stretch (PSS) test and simple shear (SS) test have been designed to explore a wider deformation space i.e., different combinations of shear and tensile strain in the stitch direction. In addition to these new shear tests, the PF test was performed with pure-UDNCF samples at different stitch pre-stretched levels to provide insights into unexplored areas of the (shear angle) - (tensile stretch) parameter space. The plotted results in the (shear angle) - (tensile stretch) parameter space revealed that the pure-UDNCF was sensitive to pre-stretching in the stitch direction during sample preparation. A new method for determining the amount of pre-stretch in each test was introduced, and the results were adjusted to be more realistic and accurate.

Except for the tensile test performed in the stitch direction, all the tests generated more than one contribution from the three low-stiffness deformation modes to the axial force i.e., shear, in-plane bending and tensile strain in the stitch direction. An iterative approach was developed to isolate the three stiffnesses of the pure-UDNCF by considering the combined experimental results of the PSS, SS, tensile, PF test, and pre-stretched PF tests. This has led to a new method of determining inplane bending stiffness using experimental data. The separate contributions from shear, in-plane bending and tensile strain in the stitch direction, provided good predictions of the measured axial force of each of the experiments. In addition to the shear tests, a cantilever bending test was performed on pure-UDNCF to measure the out-of-plane bending stiffness of the fabric. Furthermore, a new experiment was developed to measure the in-plane bending of pure-UDNCF qualitatively. This test could be further developed to quantify in-plane bending stiffness and is flagged as future work. Appropriate data obtained by these experimental methods can be used as input parameters to constitutive models to predict the material behaviour.

Following the shear and bending characterisation tests, hemispherical forming experiments were performed. Two fixation methods, i.e., acetone/epoxy solution and adhesive spray, were used to maintain the desired shape and facilitate the handling of formed specimens during post-analysis. Fabrics with different tow-stitch orientations were used to form the mono and bilayer hemispherical specimens. A combination of a 3D laser scanner (Escan H) and 3D modelling software (Autodesk 3ds Max) was successfully used to determine the shear angles and stretching in the stitch direction at selected locations on the formed hemispheres. Post-analysis of the formed bilayer hemispheres revealed that both fixation procedures provided long-term stability; however, the acetone/epoxy method caused fewer defects than the adhesive spray method due to facilitating inter-ply sliding and reducing the internal stresses between the two plies. The adhesive spray method worked well in modifying the fabric forming behaviour, offering improved control over fibre orientation, and helping to reduce gaps or inconsistencies in the fabric. Therefore, the fast and simple adhesive spray method can optimise the forming behaviour and be used locally to modify the behaviour in specific locations. These experimental forming results provide a solid foundation for validating numerical models that predict the forming behaviour of pure-UDNCF.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Additional Information: Supported by funding from the National Manufacturing Institute Scotland (NMIS) and the Scottish Research Partnership in Engineering (SRPe).
• Subjects: T Technology > TS Manufactures
• Colleges/Schools: College of Science and Engineering > School of Engineering
• Funder's Name: National Manufacturing Institute Scotland (NMIS), Scottish Research Partnership in Engineering (SRPe)
• Supervisor's Name: Harrison, Dr. Philip, McCarthy, Dr. Eddie and Zhang, Dr. Mingfu
• Date of Award: 2025
• Depositing User: Theses Team
• Unique ID: glathesis:2025-85018
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 08 Apr 2025 15:49
• Last Modified: 09 Apr 2025 08:06
• Thesis DOI: 10.5525/gla.thesis.85018
• URI: https://theses.gla.ac.uk/id/eprint/85018

Related URLs

• Article DOI