The building environment has high energy consumption and carbon emission which contributes to the rapid climate change. The changing climate condition not only impacts the natural environment but also our habitation. To improve the building performance in both energy and indoor comfort, performance-based design is commonly applied in architectural design. In general, a building lifespan is around 80 years, which means the buildings are designed and constructed now would face different climate conditions than the weather we used to evaluate the performance. To prevent the influence caused by climate change, this research develops computational workflows to generate predicted future climate conditions and multi-objective optimization to study the relationship between design decisions, climate conditions, and building performance. The conclusion of this research is based on the comparison and discussion of the result of optimization from five different climate and design scenarios with a single case study building.

One result of climate change is the increasing strength and frequency of wind events. This creates a problem for the also increasing number of high-rise buildings many of which are of unconventional shape. However, current methods for calculating wind response either do not account for these geometries, such as the Eurocode or are prohibitively expensive and time-consuming, such as physical wind tunnel tests. This thesis aims to address this issue by developing a computational method by which one can analyse the structural effects of wind on a building and optimise the external geometry to reduce those effects in the early design phase. The method involves the combination of three main algorithms: Computational Fluid Dynamics (CFD) to simulate the wind and the pressure it exerts on a building, Finite Element Analysis (FEA) which calculates the structural effects such as deflection and stresses due to these forces, and an optimisation algorithm which can iteratively manipulate an input geometry to obtain better performance. For this thesis, a tool based on the method was developed in Grasshopper, the visual scripting plugin for Rhinoceros3D. Existing plugins were used for the main algorithms while custom scripting was used to combine them into a single tool that was made relatively easy to use and returned quick results.
The methodology involved extensive research into the various aspects of the method. This was followed by the development of the method throughout which testing and validation were performed to determine its accuracy and timeliness. Case study buildings were tested with the goal of reducing structural material use. In all tests, the mass of structural material needed was reduced by allowing the optimisation algorithm to manipulate only the external geometry of the building. This produced a tool within Grasshopper and a set of guidelines for developing such a method.