Electronic Thesis and Dissertation Repository

Degree

Doctor of Philosophy

Program

Civil and Environmental Engineering

Supervisor

Dr. Girma Bitsuamlak

2nd Supervisor

Dr. Ashraf El Damatty

Joint Supervisor

Abstract

Wind is the governing load case for majority of tall buildings, thus requiring a wind responsive design approach to control and assess wind-induced loads and responses. The building shape is one of the main parameters that affects the aerodynamics that creates a unique opportunity to control the wind load and consequently building cost without affecting the structural elements. Therefore, aerodynamic mitigation has triggered many researchers to investigate various building shapes that can be categorized into local (e.g. corners) and global mitigations (e.g. twisting). Majority of the previous studies compare different types of mitigations based on a single set of dimensions for each mitigation types. However, each mitigation can produce a wide range of aerodynamic performances by changing the dimensions. Thus, the first millstone of this thesis is developing an aerodynamic optimization procedure (AOP) to reduce the wind load by coupling Genetic Algorithm, Computational Fluid Dynamics (CFD) and an Artificial Neural Network surrogate model. The proposed procedure is adopted to optimize building corners (i.e. local) using three-dimensional CFD simulations of a two-dimensional turbulent flow. The AOP is then extended to examine global mitigations (i.e. twisting and opening) by conducting CFD simulations of three dimensional turbulent wind flow. The procedure is examined in single- and multi-objective optimization problems by comparing the aerodynamic performance of optimal shapes to less optimal ones. The second milestone is to develop accurate numerical wind load evaluation model to validate the performance of the optimized shapes. This is primary achieved through the development of a robust inflow generation technique, called the Consistent Discrete Random Flow Generation (CDRFG). The technique is capable of generating a flow field that matches the target velocity and turbulence profiles in addition to, maintaining the coherency and the continuity of the flow. The technique is validated for a standalone building and for a building located at a city center by comparing the wind pressure distributions and building responses with experimental results (wind tunnel tests). In general, the research accomplished in this thesis provides an advancement in numerical climate responsive design techniques, which enhances the resiliency and sustainability of the urban built environment.


Share

COinS