Computational Design Optimization of Large-Scale Building Structures: Methods, Benchmarking & Applications

近年来，大型建筑的数量呈爆炸式增长，如高层建筑和大跨度屋顶（例如，世界上最高的20座建筑中，除了一座外，其余都是在过去15年中建造的，几乎所有建筑的高度都超过400米；此外，最近估计到2050年，将需要400万幢40层的摩天大楼来适应全球城市人口的增长）。然而，目前这种结构的形式通常以一种特别的方式确定，优化技术的应用非常有限，尽管这种技术现在在其他工业部门（例如汽车和航空航天）中经常使用。这意味着材料消耗和相关的温室气体排放往往会远远高于必要的水平，而允许包含节能功能（如采光井或中庭）的新型结构配置往往会被忽视。在本项目中，将专门针对大型建筑结构开发高效的数学优化方法，并用于自动识别结构单元的有效布局。这将为给定的设计概要确定“绝对最小材料参考设计”，为评估替代结构布局的相对效率提供了一种强大的新方法。还将开发自动生成更简单和更实用的结构布局的方法，这些布局消耗的材料比绝对最小数量的材料多不了多少。这些方法将用于确定一系列应用的结构有效布局，包括高层建筑外骨骼设计和大跨度顶棚设计。考虑到高层建筑，最近的发展是使用外骨骼“斜线”，它清晰地表达了结构系统，并且被认为比传统的解决方案更有效。然而，使用任何预定义的配置将隐含地抑制效率，并且可以使用该项目中开发的工具来考虑大量的可选布局。考虑到大跨度的雨棚屋顶，比如那些用于体育场馆、展览馆和工厂的屋顶，通过采用更有效的元素布局来减少材料消耗，从而形成一个“良性循环”，因为随着结构自重的减少，支撑它所需的结构材料的数量也减少了。该项目将为实践者提供实用工具和指导，并为学生提供教育材料。这项研究的成功交付有望极大地提高工程师设计结构高效的大型建筑的能力。

Recent years have seen an explosion in the number of large-scale structures such as tall buildings and long span roofs (e.g. all but one of the world's 20 tallest buildings was constructed in the last 15 years, almost all of which are over 400m in height; furthermore, it has recently been estimated that 4 million skyscrapers of 40 stories will be required by 2050 to accommodate worldwide urban population growth). However, currently the forms of such structures are usually identified in an ad-hoc manner, with very limited application of optimization techniques, despite the fact that such techniques are now routinely used in other industrial sectors (e.g. automotive and aerospace). This means that material consumption and associated greenhouse gas emissions will often be far higher than necessary, and novel structural configurations which permit inclusion of energy efficient features such as light wells or atriums will often be overlooked.In this project highly efficient mathematical optimization methods will be developed specifically for large-scale building structures, and used to automatically identify efficient layouts of structural elements. This will enable determination of the 'absolute minimum material reference design' for a given design brief, providing a powerful new means of evaluating the relative efficiency of alternative structural layouts. Methods will also be developed to automatically generate simpler and more practical structural layouts, which consume little more material than the absolute minimum quantity.The methods will be used to identify structurally efficient layouts for a range of applications, including tall building exoskeleton design and long-span canopy roof design. Considering tall buildings, a recent development has been the use of exoskeleton 'diagrids', which give a clear expression of the structural system, and are perceived to be more efficient than conventional solutions. However, the use of any predefined configuration will implicitly inhibit efficiency and vast numbers of alternative layouts will be able to be considered using the tools to be developed in this project. Considering long-span canopy roofs, such as those used in sports stadia, exhibition halls and factories, reducing material consumption by adopting a more efficient layout of elements leads to a 'virtuous circle' since as structural self-weight is reduced, so does the amount of structural material required to support this.The project will result in the development of practical tools and guidance for practitioners, and educational materials for students. Successful delivery of the research can be expected to dramatically improve the ability of engineers to design structurally efficient large-scale buildings.