Location
London
Event Website
http://www.csce2016.ca/
Description
Wind is one of the governing load cases for tall building design, which produces high level of straining actions, deflections and lateral and transverse vibrations. Keeping those vibrations within the comfort limits is becoming a key aspect in tall building design, especially for buildings with high aspect ratio. Improving the aerodynamic performance of the tall building by modifying its shape can lower building motions, which reduces the additional expenses for external damping systems and alleviate the high cost associated with lateral support systems. In the present study, an aerodynamic shape optimization procedure is developed by combining Computational Fluid Dynamics (CFD), optimization algorithm and Artificial Neural Network (ANN). The developed procedure utilizes ANN as a surrogate model for evaluating aerodynamic properties, which is pre-trained using two-dimensional CFD analysis. The current study investigates the validity of the developed procedure by conducting a high accuracy, three-dimensional Large Eddy Simulation (LES) based analysis on the optimal building shapes. It was observed that utilizing two-dimensional CFD simulations in the optimization procedure can help identify effective cross-sections of tall buildings.
Included in
NDM-530: AERODYNAMIC OPTIMIZATION TO REDUCE WIND LOADS ON TALL BUILDINGS
London
Wind is one of the governing load cases for tall building design, which produces high level of straining actions, deflections and lateral and transverse vibrations. Keeping those vibrations within the comfort limits is becoming a key aspect in tall building design, especially for buildings with high aspect ratio. Improving the aerodynamic performance of the tall building by modifying its shape can lower building motions, which reduces the additional expenses for external damping systems and alleviate the high cost associated with lateral support systems. In the present study, an aerodynamic shape optimization procedure is developed by combining Computational Fluid Dynamics (CFD), optimization algorithm and Artificial Neural Network (ANN). The developed procedure utilizes ANN as a surrogate model for evaluating aerodynamic properties, which is pre-trained using two-dimensional CFD analysis. The current study investigates the validity of the developed procedure by conducting a high accuracy, three-dimensional Large Eddy Simulation (LES) based analysis on the optimal building shapes. It was observed that utilizing two-dimensional CFD simulations in the optimization procedure can help identify effective cross-sections of tall buildings.
https://ir.lib.uwo.ca/csce2016/London/NaturalDisasterMitigation/22
