Thesis
Out-of-autoclave processing development of aerospace carbon fibre polymer composites
- Creator
- Rights statement
- Awarding institution
- University of Strathclyde
- Date of award
- 2026
- Thesis identifier
- T17663
- Person Identifier (Local)
- 202088883
- Qualification Level
- Qualification Name
- Department, School or Faculty
- Abstract
- Composite materials, particularly carbon fibre-reinforced polymers (CFRPs), have garnered significantattention in the aerospace industry due to their lightweight and high-strength properties. Autoclaving,a well-known method for producing high-quality CFRPs suitable for aerospace applications, faceschallenges such as high energy consumption making it less suitable for high-volume production. ResinTransfer Moulding (RTM) has emerged as a viable alternative, offering advantages over autoclaving,including lower temperature latency during curing stage. This potentially can reduce the likelihood oftemperature overshooting within thick CFRP laminates and result in a more uniform temperaturethrough part thickness, consequently lowering residual stress.This PhD thesis investigates and optimises the through-thickness temperature of thick CFRP laminatesduring curing stages by studying the effect of a wide range of manufacturing parameters for RTMprocessing. This was carried out by developing a thermal and cure simulation tool usingMATLAB_Simulink and connecting the model to the response surface method (RSM).Initially, cure kinetics of epoxy resin were studied through thermal analysis. As a result of the analysis,a deceleration reaction was detected and described by the Sestak-Berggren Kinetic model. The totalheat of reaction, activation energy, pre-exponential factor, and n and m constants were measured byfollowing the Kissinger-Akahira-Sunose (KAS) and Malek’s methods to verify that the selected modelcan comply with the epoxy resin. The comparison of the reaction rate obtained from the cure kineticequation with the thermal analysis showed close agreement.The CFRPs were manufactured using vacuum bagging mainly due to the simplicity of embeddingthermocouples inside the mould and between the carbon fibre plies. Laminates with 10, 20, and 30mm in thickness were manufactured while their temperature and degree of cure at the top, middle,and bottom of the lay-up were recorded.Void content, density, fibre weight fraction, in-plane and through-plane thermal conductivity (TC) ,and heat capacity of the manufactured samples were measured through different methods andshowed uniformity between all samples.The RSM was used to identify the impact of different variables such as the thickness of the part,heating cycle, one-sided and two-side heating conditions, thermal conductivity of the mould and thecomposites, and ambient temperature on the temperature distribution inside the CFRP laminates.Eight parameters (response variables) were defined in the RSM to give a clear picture of thetemperature and degree of cure within the parts. Maximum temperature (T_max), maximumtemperature difference through the thickness (?T_max), and location of the maximum temperature (“C”) are among response variables that were studied for their dependence of other controllablevariables. The factor “C” is helpful in detecting curing patterns through the part thickness, which islinked to a reduction in the level of induced stress during the curing process. Inside-to-outside curingpattern should happen before resin gelation points (preferably from the start of the curing) to beeffective in reduction of residual stress.The RSM with central composite design (CCD) was used to create more than two thousand test runsrequired to detect the interaction between the manufacturing variables (controllable variables) andresponse variables. This was carried out by developing a simulation tool that could predict thetemperature and state of the cure through the CFRP laminates under desired heating condition. A 1Dthermal model based on the finite differential method (FDM) was developed in MATLAB_Simulink,and the experimental results verified its performance.The influential variables and their effect on the eight response variables were investigated fordifferent scenarios. The mould with low thermal conductivity resulted in a smaller ?T_max and T_max,specifically for thick laminates. For composites with lower thicknesses, the low thermally conductivemould showed little benefit and have result in longer curing time.The results from varying ambient temperatures demonstrated that selecting an appropriate ambienttemperature can significantly reduce the temperature difference through the thickness of the part(?T_max), as well as minimize the temperature overshoot (T_max) and variations in the degree ofcure. Notably, ambient temperature was found to be the most influential factor in determining thelocation of T_max, effectively shifting the curing behaviour to inside-to-outside pattern. This shiftsuggests that with a well-controlled ambient heating program, one-sided heating can achieve similarthermal and curing uniformity as two-sided heating. However, unlike two-sided systems, one-sidedheating requires less complex tooling and setup. This can lead to significant reductions inmanufacturing cost, tooling complexity, and energy consumption, particularly in large-scale or thickcomposite parts.The optimal ambient temperature and heating schedule were developed using Response SurfaceMethodology (RSM), enabling precise control over curing conditions and ensuring high-quality resultswith a more cost-effective setup.Overall, the combination of the RSM and Simulink built a comprehensive model that was able topredict the temperature and degree of cure in composite laminates. The model was also used tooptimize the process by suggesting the correct values for the manufacturing parameters to providethe response variables close to the target defined by the user. This was confirmed by setting the CFRP curing conditions according to the model-defined parameters and comparing the temperature values measured at the top, middle and bottom of the samples with the simulation. The model showed apromising performance in designing manufacturing conditions and a potential to be used on morecomplicated geometries. The developed model can be adjusted for a wide range of reinforcedcomposites and different resin systems, although it remains subject to further testing and validation.
- Advisor / supervisor
- Yang, Liu
- Resource Type
- DOI
- Date Created
- 2025
- Funder
Relations
Items
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
|
PDF of thesis T17663 | 2026-03-25 | Public | Download |