Optimising the geometry of 3D printed steel structures against local and global instability.

研究的首要目标是研究如何通过改变其表面的形式来提高钢截面的屈曲强度，以类似于等效棱柱的屈曲模式。作者之前已经对这一问题进行了研究，在纯压缩条件下对修改后的RHS短柱进行了数值试验（Chater & Wang，2019），并对EAS截面进行了实验室试验。本博士旨在扩展这项工作如下：1）在各种负载条件下进行测试，以达到TRL 4。这将采取在与现有结构的实际条件相当的情况下对所使用的技术进行若干模拟的形式。这可能包括4个部分：在弯曲、剪切、轴向和组合载荷下测试相同的短柱。需要做进一步的研究，以确定最佳做法，这些不同的负载，以最好地捕捉的行为的存根，这将是一个文献综述的一部分。测试的一个较长的长度的柱（和/或梁）在各种负载条件下，以建立已知的局部失效机制和全局故障之间的关系。同样，这到底是什么将取决于对现有文献的回顾。在直接取自现有建筑设计的较大结构系统中对这些部分进行测试。可以与外部顾问合作，以确定合适的测试系统。疲劳测试和生命周期分析，以确保使用这些部件不会出现意外的长期问题。2）测试相同立柱的修改版本，充分利用增材制造的能力，改变整个截面的材料厚度。这些部分的优化可能由人工神经网络（ANN）驱动，因为这些网络广泛用于当今的非线性行为优化。本节的目标是找到具有给定材料量的单个完全连接的连续元件可实现的最大屈曲能力。3）测试传统轧制后冷压成先前定义的拓扑结构的截面。这样做的目的是使该技术更容易获得，通过将其与增材制造相关的生产率，成本和环境影响的限制分开，将其扩展到更多的受众。在所有这些情况下，将通过ABAQUS中的数值分析进行测试，并在可能的情况下进行实验室测试，以确保这些测试的有效性。在所有上述内容之前，将进行广泛的文献综述，旨在找到最佳实践和先例，以最好地实现所概述的研究目标。

The overarching goal of the research is to study how the buckling strength of steel sections can be improved by changing the form of their surfaces to resemble buckling modes of equivalent prismatic columns. Previous research on this matter has already been carried out by the author, with numerical tests done on modified RHS stub columns under pure compression (Chater & Wang, 2019), and lab tests carried out on EAS sections. This PhD seeks to expand on that work as follows:1) Testing under a wide variety of loading conditions to achieve TRL 4. This would take the form of several simulations of the technology used within situations comparable to actual conditions found in existing structures. This would probably consist of 4 parts:Testing of identical stub columns under bending, shear, axial and combined loads. Further research needs to be done to identify best practices for these varied loads in order to best capture the behaviour of the stubs; this would be a part of the literature review.Tests of a longer length of column (and/or beam) under a variety of loading conditions to establish the relationship between known local failure mechanisms and global failures. Again, exactly what this constitutes will be dependent on a review of existing literature. Tests of these sections within larger structural systems taken directly from existing building designs. Collaboration may be undertaken with external consultants to identify suitable systems in which to test this.Fatigue tests and lifecycle analysis, to ensure no unexpected long-term issues from use of these parts.2) Testing of modified versions of the same columns, which fully utilise the capabilities of additive manufacturing to vary the thickness of material throughout the section. Optimisation of these sections will likely be driven by an artificial neural network (ANN), as these are widely used in optimisation for non-linear behaviours in the present day. The goal of this section would be to find the greatest buckling capacity achievable by a single, fully connected continuous element with a given amount of material.3) Testing of sections cold-pressed into the previously defined topology after being rolled traditionally. The aim of this is to make the technology more accessible, extending it to a greater audience by separating it from the limits on production rate, cost and environmental impacts associated with additive manufacturing. In all of these cases, testing will be carried out both via numerical analysis in ABAQUS, with lab tests following after wherever possible to ensure the validity of these tests. Prior to all of the above would be an extensive literature review, aiming to find best practices and precedents for how best to approach the outlined research goals.