Cellular structures have a wide variety of applications from impact absorption in packaging to structural sandwich panels in aerospace applications. The mechanical properties that these structures maintain with a vastly reduced mass over solid materials makes them desirable in weight sensitive applications. Honeycomb seen in corrugated cardboard and composite core materials is a widely adopted cellular structure, however, its highly anisotropic material properties limit its use to Two Dimensional (2D) loading cases where geometric complexity is low. Foams can be used to fill the space where complexity is high or loading direction is variable. However, the advantages gained by an isotropic structure are balanced with lower maximum properties. The current state of art has evolved to engineered cellular structures with increased mechanical properties over foams while remaining isotropic. Additive manufacturing has been adopted to produce the complex cell-level geometry but comes with the side-effect of increased cost and limited part volume.

In this work, the manufacture of Triply Periodic Minimal Surfaces (TPMS), a family of highperformance cellular structures is explored. The use of low-cost desktop Material Extrusion (MEX) additive manufacture is proposed to reduce the manufacturing cost and size constraints of existing methods that employ polymer or metal Powder Bed Fusion (PBF). Limitations are identified in the generation of Three Dimensional (3D) models used for the printing process. A cookie-cutter method to compile many unit cells and cut out a part of desired geometry is developed. To further optimise the model generation process where hundreds or thousands of cells are used, a method using surface meshes is used and a proof of concept slicer is developed to allow the geometry to be exported directly to g-code and printed without the need for any 3D model that is limited by computer processing power or file size.

Low cost MEX 3D printing is known for its low resolution and anisotropic mechanical properties. Amethod is developed to produce samples of cellular structure of high enough quality and repeatability to be used in experiments and real-life applications. Compression testing of Schoen Gyroid and Schwartz Diamond TPMS is performed employing the MEX 3D printing manufacturing techniques. Samples of 10%, 15% and 20% Relative Density (RD) are compressed past their densification point to determine the linear elastic and plastic deformation response as both the structural and energy absorbing properties are important. The mechanical properties are found to be of similar magnitude to that of existing literature using other manufacturing techniques. With loading of samples both parallel and transverse to the printed layers the anisotropy of MEX printing is shown to have a large impact in the anisotropy of printed structures. When loaded parallel to the layers, a reduction in stiffness, strength and energy absorption is observed along with a change in failure mode.

Finite Element (FE) analysis is performed on the unit cell of the Schoen Gyroid, Schwartz Diamond and Schwartz Primitive TPMS structures using an optimised method of mesh generation and analysis to reduce the computational requirement and therefore analysis time. Building on existing methods, a process is developed to efficiently create surface meshes of the TPMS structures with high quality elements. A homogenisation method generally used for the analysis of composite laminates is adapted to work with surface cellular structures. The elastic properties of the three TPMS structures is obtained in an automated fashion that enables the rapid analysis of many structures or parameters. It is shown that comparable results can be achieved with vastly reduced mesh sizes when the shell models are compared to literature using continuum elements.

With the anisotropy of MEX printing shown to influence the anisotropy of the cellular structures, a method to improve the layer adhesion of MEX printed parts is explored. A method of staggering neighbouring extrusions of plastic in the Z-plane is hypothesised to reduce the porosity of the finished part, increase the contact surface area between extrusions and allow the addition of more plastic; overall increasing the inter-layer bonding and therefore reducing the anisotropy. The staggering method visually improves the sample quality with less displacement of molten plastic when the Extrusion Multiplier (EM) is increased. Staggered samples are shown to have a lower porosity than non-staggered with all samples having a decreasing porosity with increasing EM. Unfortunately, in an initial phase of experiments staggering showed little improvement in mechanical properties with the biggest increase seen with the impact strength of PolyLactic Acid (PLA) samples while staggered Acrylonitrile Butadiene Styrene (ABS) samples showing overall lower performance. However, a second phase of refined experimental testing showed an increase in mechanical properties across all samples. ABS samples showed particular improvement in repeatability and a larger increase in mechanical properties compared to PLA samples. Further comparison to samples printed using the natively-sliced methods indicate that the concentric tube method of implementing the staggering artificially reduced the mechanical properties of both the staggered and non-staggered control samples. Staggering has therefore been been demonstrated to reduce porosity and, in some cases, to improve strength; further work will be needed to conclusively prove improved mechanical properties.

A case-study for the application of functionally graded TPMS structures is explored using an optimised hard-point in a composite sandwich panel. A 2D axi-symmetric model is developed in ABAQUS to analyse the relative deflection, mass and stress concentration of three hard-point implementations. These are a conventional solid hard-point, a graded hard-point and no hard-point. The graded hard-point was shown to have a acceptable level of deflection when compared to the unacceptable no-hard-point and the regular solid hard-point. A reduction in mass and an improved stress distribution in the composite panel was also observed. A method of generating the 3D model used to manufacture a graded hardpoint is explored. The method creates a variation in relative density by changing the wall thickness in 3D space. A Stereolithography (.STL) file is generated for 3D printing using the Grasshopper plugin for the Rhino Computer-Aided Design (CAD) system. The limitations of MEX printing for the manufacture of this file is discussed.