The Analysis of Process-induced Deformations in Carbon-Fibre Reinforced Poly-ether-ketone-ketone (CF/PEKK) Composite Structures during the Thermoforming Process

Abstract

Process-induced Deformations (PIDs), such as warpage of flat structures and spring-in/ spring-out of curved structures, are some of the major manufacturing process challenges associated with the aerospace composite industry. These PIDs cause dimensional instability, in which the final structure dimensions deviate from the design requirements and accordingly affect the structure quality, leading to part rejection and scraping of the structure due to assembly stage issues. In this dissertation, a numerical-experimental approach is adopted to predict the spring-in of manufactured CF/PEKK composite structures using a proposed three-network model based on the time-, temperature-, and crystallinity-dependent relaxation response during the thermoforming process in addition to investigating warpage behaviour experimentally. The effect of design and processing conditions on warpage phenomena is investigated under non-isothermal forming conditions, using a multi-level factorial design of experiments and the analysis-of-variance (ANOVA), to help reduce the warpage effect in manufactured composite panels. The effect of the processing conditions at melt state and cold crystallization, and consequentially the degree of crystallinity, on the viscoelastic and interfacial properties and relaxation behaviour is examined using Differential Scanning Calorimetry (DSC), Dynamic Mechanical Analysis (DMA) interlaminar shear strength (ILSS) tests. The effect of time, temperature and degree of crystallinity on the relaxation behaviour of thermo-rheological complex material (TCM) is investigated using the Time-Temperature-Crystallinity Superposition Principle (TTCSP) applied on sets of multi-temperature DMA relaxation tests to generate great-grand master curves. The Three Network Model (TNM), a thermomechanical model, is calibrated using these relaxation test data in MCalibration software and the validated material model, consequentially, is then implemented as a user-subroutine UMAT in ABAQUS to predict process-induced deformation during the thermoforming simulation of composites structures. The calibrated TNM model showed a good prediction of the relaxation behaviour of PEKK composite at a high-temperature range between the crystallization temperature and glass transition temperature with an average error of 16% for samples consolidated using a single-hold cycle. The thermoforming simulation results have shown a 6% agreement with the measured process-induced deformation in magnitude only, as the FE outcome predicted a spring-back whereas the expected deformation was spring-in. The FE model can predict the magnitude using the calibrated TNM model, however, further assumptions need to be considered in simulations to achieve the expected spring-in deformation. The outcomes of the dissertation will help design engineers, material and process engineers and tooling engineers in predicting the relaxation behaviour of consolidated composite structures, in addition to estimating the post-manufacturing deformations of thermoplastic composite structures at the early stages of the product cycle, which will save material cost and time, allowing them to make early decisions in regards to composite lay-up, tooling shape, and consolidation cycle parameters.

Date of Award	15 Dec 2023
Original language	American English
Supervisor	Rehan Umer (Supervisor)