Echeta, Ifeanyichukwu
(2023)
Analysis of defects in additively manufactured lattice structures.
PhD thesis, University of Nottingham.
Abstract
Additive manufacturing is a popular area of research because it greatly increases design opportunities, allowing for significantly more geometric freedom than in more established manufacturing methods, such as machining, casting and forming. A relatively small set of additive manufacturing processes are consistently used for the manufacturing of lattice structures, and these processes produce characteristic defects and geometric deviations within lattice structures.
In this thesis, a modelling approach is presented for the generation of surface models of strut-based lattice structures into which defects and geometric deviations can be added. Conversion of the surface models into tetrahedral meshes for finite element (FE) analysis is also demonstrated. Signed distance functions (SDFs) form the foundation of the model and can be used to create surfaces of ideal lattice structures. The thesis demonstrates how modification of the signed distance function allows for the inputting of geometric deviations—namely, waviness, radius variation and elliptical cross sections. Surface defects are modelled by defining an additional function that applies displacements to the surface produced by the signed distance function. To understand the limitations of the proposed modelling approach, a sensitivity study is performed wherein the underlying parameters of the approach are modified to observe their impact on three quantities: SDF error, meshing error and mesh quality.
X-ray computed tomography (XCT) was used for obtaining original data on geometric deviations and surface defects in lattice structures, more specifically, a BCCZ lattice structure. Cross sectional measurements of the struts was performed, as well analysing the strut surfaces to observe locations of increased surface defects. Comparisons were made between the design’s vertical struts and inclined struts. The XCT results showed the inclined struts to be significantly more prone to geometric deviations; radius variation, waviness and texture bias all showed greater deviations in the inclined struts. The cross sectional data, grouped by strut orientation, was fitted to probability density functions (PDFs) which were used in subsequent stages for generating lattice structures with geometric deviations statistically equivalent to the XCT measurement.
The BCCZ lattice structures were also subjected to compression testing for determining the Young’s modulus of the design, which was determined to be 984.1 MPa. The proposed modelling approach was then configured, using the PDFs derived from the XCT data to generate a model of a lattice structure with geometric deviations applied. Upon the application of the geometric deviations, the simulated Young’s modulus reduced from 4148 MPa to 4023 MPa, suggesting that the introduction of geometric deviations does indeed reduce stiffness, however, these results are a significant overestimation of the experimentally determined Young’s modulus. A number of areas could be explored to improve this disparity, in particular, the updating of the material model used in the analysis.
In summary, the work in this thesis demonstrates the versatility of SDFs for the modelling of strut based lattice structures. The XCT results showed strong trends between strut overhang angle and the exacerbation of geometric deviations and surface defects. The cross sectional data from the XCT measurement was well described by the PDFs; the simulated data showed very strong agreement to the XCT data. The FE modelling requires further investigation to improve its agreement with experimental data.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
