THESIS
2016
xx, 223 pages : illustrations (some color) ; 30 cm
Abstract
Nowadays, a cost-effective and speedy tall building design is important for any engineering
company to stay competitive. Due to fast-track project and higher labor cost, time and cost of
a wind tunnel experiment can be impractical when there is a shape change or unavailable
laboratory during the design phase. Because the efficiency of time transient CFD simulations
is limited by restricted mesh and time steps, there is an increasing demand to develop an
effective numerical wind load simulation.
Although RANS simulations are computational inexpensive to predict mean wind loads, the
lack of fluctuating pressure makes RANS inapplicable for wind load prediction. In this research,
a hybrid simulation is developed to utilize the results of RANS simulation and implement a stochasti...[
Read more ]
Nowadays, a cost-effective and speedy tall building design is important for any engineering
company to stay competitive. Due to fast-track project and higher labor cost, time and cost of
a wind tunnel experiment can be impractical when there is a shape change or unavailable
laboratory during the design phase. Because the efficiency of time transient CFD simulations
is limited by restricted mesh and time steps, there is an increasing demand to develop an
effective numerical wind load simulation.
Although RANS simulations are computational inexpensive to predict mean wind loads, the
lack of fluctuating pressure makes RANS inapplicable for wind load prediction. In this research,
a hybrid simulation is developed to utilize the results of RANS simulation and implement a stochastic Kinematic Simulation to generate a wind flow field that satisfies a specified energy
spectrum. Due to the efficiency of RANS and KS, the hybrid simulation can realize the potential
of full-scale wind load simulations on buildings.
Given predicted mean and fluctuating pressure from hybrid RANS-KS simulation, spatial
correlations between different locations on a building are needed to integrate the results of
hybrid simulation into wind loads for building design. A great advantage using wind tunnel
tests is the ease of obtaining pressure measurements for various building shapes after an initial
setup. Based on a well-defined experimental database, an experimental database model using
Radial Basis Function is applied to estimate the correlation matrix of fluctuating pressure. The
physical interpretations of correlation matrices between faces are studied and implemented.
This correlation predictive model is integrated with hybrid RANS-KS simulation to predict the
design wind load in a reliable and cost-effective manner.
Based on the results of spatial-temporal wind pressures, understanding the relationship between
time varying pressure distributions and wind loads is a practical challenge. A statistical-based
Load Categorized Representation is developed to investigate the crucial statistical identifiable
load characteristics for the structural design of lateral load resisting system without prior
knowledge. Utilized statistical independence characteristics of Independent Component
Analysis, the statistical model summarize the load characteristics on a building by some concise
load oriented representations in both spatial and temporal aspects. These representations can
also be adapted to compare the load characteristics of different buildings or wind conditions.
Numerous examples are presented to demonstrate the efficiency and practicality of proposed
algorithms in numerical simulation and characteristic representation of wind loads on buildings.
Post a Comment