2023 Theses Doctoral
Numerical modeling of non-stationary boundary-layer flow in urban areas
The atmospheric boundary layer (ABL) has been the subject of active research due to its relevance to human welfare. Among the various research approaches, large eddy simulation (LES) has become a prevalent tool for studying ABL flow phenomena because of its inherent ability to provide reasonably accurate and detailed spatially-distributed velocity information without the need for many ad-hoc tuning parameters.
Motivated by the need to understand fundamental ABL processes in their simplest form, previous LES investigations have primarily focused on idealized flow regimes characterized by, e.g., statistically stationary or homogeneous flow conditions. However, these simplified scenarios fail to capture much of the complexity that characterizes the real-world ABL, including its impact on various applications. Thus, there is a pressing need to explore more realistic flow scenarios that can capture a broader range of processes that occur in the ABL.
This dissertation focuses on two aspects of real-world ABL flow complexity: non-stationary effects and flow in complex terrain. The first part of this work examines how non-stationarity—a prevalent but long-overlooked characteristic of real-world ABL flows—impacts the transfer of momentum between the land surface and the atmosphere. Specifically, the focus is on flow driven by time-varying pressure gradients, which are a common source of flow unsteadiness. The analysis is based on LES of pulsatile flow over resolved urban-like surfaces. This study proposes a sensitivity analysis to explore how key parameters governing ABL non-stationarity affect flow dynamics. Results shed light on the mechanisms responsible for changes in momentum transport and turbulence generation under non-stationary conditions. Based on these findings, two phenomenological models are proposed to capture variations in aerodynamic parameters under such conditions. The same flow system is further characterized from a coherent structure perspective to identify fundamental mechanisms controlling the transport of momentum. Non-stationarity is found to yield a time- and space-varying shear rate, which directly affects the topology of dominant coherent structures. These structures, in turn, modulate the ejection-sweep pattern, which is the dominant mechanism for momentum transfer in the ABL.
The second part of the dissertation is devoted to evaluating the predictive abilities of finite-volume (FV) solvers in LES of ABL flows. The motivation is that these solvers are naturally suited for studying turbulent flows in complex terrain geometry, but a full assessment of their performance is still lacking. A suite of LES of turbulent channel flow are conducted at a moderate Reynolds number (??? = 2000), using a general-purpose, second-order-accurate FV solver. Results are compared against those from a mixed pseudo-spectral and finite-differences solver and the direct numerical simulation (DNS) benchmark, with a lens on first- and second-order statistics, and their sensitivity to the choice of numerical and physical parameters. Based on the findings, the study recommends an optimal setup for ABL simulations based on FV solvers.
Files
-
LI_columbia_0054D_17819.pdf application/pdf 4.1 MB Download File
More About This Work
- Academic Units
- Civil Engineering and Engineering Mechanics
- Thesis Advisors
- Giometto, Marco Giovanni
- Degree
- Ph.D., Columbia University
- Published Here
- May 10, 2023