The IsoTruss is an interesting continuous fibre reinforced polymer composite design which is closely related to open lattice composite structures, although it is currently aimed mostly at civil applications. While it currently lacks applicability to the aerospace sector, its prod
...
The IsoTruss is an interesting continuous fibre reinforced polymer composite design which is closely related to open lattice composite structures, although it is currently aimed mostly at civil applications. While it currently lacks applicability to the aerospace sector, its production process is easier to automate (or change the shape input parameters on the fly) than open lattice structures, which rely on a fixed mandrel. Although automated compaction of these structures is a challenging problem, the structure can be optimised for any (static) load case if their production process could be automated, providing a range of applications where aerodynamic contour is not a requirement. To bridge the gap between the IsoTruss and other composite lattice structures, a conceptual design for an IsoTruss derivative, termed the "iso-truss", was envisioned. This raised the following research question: Can composite continuous fibre reinforced iso-truss structures be produced using a scalable and cost-effective automated manufacturing method? Through this thesis project, an affirmative answer to this question was sought. The focus of this thesis was on the investigation of all the aspects necessary for a production of this iso-truss structure, focusing on a process suited for later automation. Initial concepts using thermoset composites like the original IsoTruss could not solve the issue of automated compaction, as these would always rely on vacuum bagging to minimise the void content within the composite members. Automated vacuum bagging of the finished iso-truss structure was considered to be an infeasible solution due to the product's complexity. Through the use of thermoplastic composite materials however, the process steps could be split up and compaction could be moved to the very beginning of the process in the form of pultrusions. The shape of the resulting iso-truss was designed to make use of this process, providing a parameterised structure which could easily be optimised. The structure would make use of straight pultrusion rods, acting as the longitudinal members in the truss, and curved pultrusion rods, which would act as helical members. The iso-truss structure would be created by joining these members together at intersections by means of thermoplastic welding, forming a continuously produced open lattice tubular structure. With this, conceptual production of an iso-truss was based on three key steps: pultrusion, forming, and intersection welding. The pultrusion process was outsourced to an outside company, vDijk Pultrusion Products (DPP), due to its high required equipment cost. It was not the aim of this thesis to re-invent this process, as it has already reached a sufficient state of maturity. Nevertheless, the process has not yet been optimised for thermoplastic based composites, leaving room for improvement in this area. Materials for the concept were based on availability. Elium resin (a PMMA based thermoplastic polymer developed by Arkema) was chosen for its similarity to existing process materials and its direct availability from DPP. As reinforcement, a Toray T700SC type carbon fibre was used. By taking pultrusion out of the research scope, forming and welding were left as processes to demonstrate. A separate welding test was isolated to provide a way to perform standardisable tests. Simultaneously, a forming process demonstrator was designed, as well as a welding jig to demonstrate assembly of the iso-truss. For the execution of the welding and forming experiments, a single batch of material was ordered from DPP. The batch was the result of a first successful attempt to make rods of this diameter with this combination of materials. Upon arrival of the samples, several tests including SEM, TGA and DSC were performed to estimate important composite material parameters. From these tests, the fibre volume fraction could be estimated to be around 70%, while the onset of T_g was estimated to be 95degC by definition of maximum loss modulus. Additionally, several interesting phenomena were observed at elevated temperatures in the form of fibre kinking when bent and circumferential decompaction when twisted, indicative of a weak fibre-matrix interface. More indications of this limited fibre-matrix interface were obtained from SEM images, some of which showed small voids around fibres. DMA tests showed a steady decline of the material's (shear) stiffness over a large temperature range, including before the onset of glass transition. This could likely be attributed to the amorphous nature of the polymer. For the execution of joining experiments, a standard intersection layout was established, after which several joining techniques were considered to create these intersections. From these techniques, heated mould welding was selected as the best currently feasible option. After performing numerous experiments to fine-tune the method of heated mould welding, a standard setup could be designed around this technique. This setup addressed methods for heating, alignment of members to be welded, a compaction mechanism, physical test preparation and physical testing. Simultaneously, a detailed design of a helical winder demonstrator was made, which was prepared for production by DEMO. The design focused on the demonstration of the mechanical aspect, leaving heating as a secondary concern. Similarly, a more detailed concept of an assembly demonstrator was discussed and designed, although a production-ready design did not prove feasible within the time frame of this project. The main goal of the intersection welding experiments was to optimise the joint strength by optimising the processing conditions. Due to the limited resin volume content, PMMA foil and epoxy adhesive were used to locally increase resin volume, greatly improving the bond strength compared to initial samples. Using mould temperatures in between 180 to 200degC, a large degree of deformation could be achieved to maximise the joint area, creating mostly consistent intersection samples. Physical tests proved that the PMMA-based joints were very fragile, while epoxy based joints performed significantly better. Estimated shear strengths for the samples varied from 3MPa for some of the lower performing PMMA-based intersections which were believed to have sustained prior damage, to as high as 13MPa for the epoxy based intersections. For the buckling tests, maximum loads in between 2.1 to 3.2kN were observed. In the final buckling test setup, diagonal members did not increase the buckling load by forcing higher-mode buckling. Instead, they offered some support in post-buckling by resisting out-of-plane deformation. Failure modes in both shear and buckling tests were within the adherend, occurring as a combination of fibre break-out, member kinking and member splitting, more indications of a poor fibre-matrix interface. In summary, the experiments proved that the investigated welding method is feasible if the base material can be improved. As of writing, such improvements have already been reported by DPP through increased process stability and (likely) better fibre sizing. For the helical winding experiments, the main goal was to demonstrate the production process of heated forming of initially straight pultrusion samples into helices. The physical test setup was designed to offer continuous support of the members over their full length, realise a gradual increase in radius of curvature and allow torsional and axial forces to be introduced independently. For the resulting setup, heating of the central cylinder and introduction of heat through both contact and a heated air chamber proved most feasible within the time frame of this project. It did not prove feasible to produce carbon-Elium composite helices due to the current material limitations, that is the fibre kinking observed in previous tests. Instead, the process was demonstrated using pure PMMA rods at forming temperatures of around 90degC, which proved highly effective. The process showed the potential to create constant curvature helices with controllable helical angle, although more testing should be performed to achieve accurate and repeatable results. It is concluded that it is theoretically possible to produce continuous fibre reinforced iso-truss structures using a scalable and cost-effective automated manufacturing method. The process would rely on existing production capabilities which can easily be scaled up or down, without greatly affecting the cost of equipment. The pultrusion, helical shaping and intersection welding processes can all be easily automated using existing techniques to be able to produce an iso-truss tube of indefinite length. The only limitation of this manufacturing method would be the variation of process parameters within the same iso-truss sample. While it is theoretically possible to vary the radius of curvature of pultrusion dies during the pultrusion process, enabling the possibility of varying the outer diameter and the intersection node-to-node distance along the length of a tube, the current state-of-the-art does not accommodate such a design. While it would also be possible to vary the cross-sectional area of each member along their length, it is considered undesirable as it would negatively affect the fibre volume content for a pultrusion-based process. Instead, it is recommended to construct an iso-truss structure in stages, being connected by a structure that is to be envisioned and produced in future research.
