Computational design optimisation is a much more powerful method than manual or trial-and-error optimisation, and it is increasingly applied to optimise structure and building envelopes. Most existing optimisation applications are only single-objective, and follow a traditional sequential process in which each subsystem or discipline is optimised in isolation in a predetermined order, assuming that the designs of other subsystems remain fixed. As a result, each subsystem optimisation in later design phases is conducted based on a single solution provided by the previous subsystem optimisation. This traditional approach leads to significant limitations including long design cycle time due to impossibility of simultaneous optimisation, and sub-optimal results due to low degree of design freedom in later design phases.
To overcome these limitations, multi-objective optimisation and an improved multidisciplinary optimisation process are increasingly promoted and become promising and active research fields in building design. This research aimed at developing a performance-based building design approach, which improves multiple performance aspects and supports trade-off decision making for the conceptual architectural design of indoor sports buildings, by taking full advantage of computational exploration and optimisation. The multiple aspects addressed included architectural requirements, daylight, energy and structural performance.
The research team, comprising the Chair of Design Informatics at TU Delft, SCUT (South China University of Technology School of Architecture) and Arup Amsterdam, explored a multidisciplinary design exploration process using statistical analysis and Design of Experiment (DoE), which is a systematic method to determine between factors that affect the design process and the outcome of the design process, rather than optimisation alone.
By using valuable design information and multidisciplinary design exploration there is potential to support reformulation of the design space and the selection of the objective function.
The multidisciplinary design exploration also has the potential to support investigation and consideration of various objective functions and allows knowledge extraction which provides crucial engineering feedback for the consideration of the design team.
In the context of green buildings and computational design, it is of great value to implement computational optimisation methods at the early design stage for improving multiple building performances across various disciplines.
The multi-objective multidisciplinary optimisation framework developed with this research overcomes the limitations of single-objective optimisation, and offers a method to make trade-off decisions between multiple (and sometimes conflicting) design objectives at the early stages of design. The incorporation of optimisation methods enhances the ability to search for high-performance solutions among a huge number of design alternatives from various disciplines simultaneously (architecture, structure, energy, and lighting under the current framework).
A typical indoor sports building project in China has been selected as a case study to test the proposed method during the research, covering the optimisations in three areas: structural design, daylight usage and energy efficiency.
With this approach, the tasks of conceptual architectural design and various building engineering disciplines can be brought together and the design approach can be shifted from a conventional, linear process, to a more collective, iterative approach. Such integrated design is central to Arup’s core vision of ‘Total Design’.
Sports buildings are a great design opportunity, but present design challenges in meeting the diverse design requirements for large-span structure with special performance requirements regarding daylight usage and energy efficiency. Multi-objective optimisation offers the potential of achieving high-performance building envelope that can enhance the medium/long term sustainability of the sports building in service.