JavaScript is disabled for your browser. Some features of this site may not work without it.

A Computational Study of Flow Over a Wall-Mounted Cube in a Turbulent Boundary Layer Using Large Eddy Simulations

Shinde, Siddhesh Dilip

Shinde, Siddhesh Dilip

2018

Abstract: Flow over a wall-mounted cube in a turbulent boundary layer (TBL) is a canonical problem with applications in many engineering systems. Atmospheric flow over buildings in an urban environment or vegetative canopies, air flow over road vehicles, flow over printed circuit boards, etc., are few examples which can be modeled by considering flow over wall-mounted cubes. Without loss of generality, the problem of interest in this work is controlling the separation region on the rear end of road vehicles to reduce aerodynamic drag. To do so, we intend to use a row of cubes placed in single line normal to the flow direction, as passive vortex generators (VGs) to reduce flow separation. Flow separation is caused by an adverse pressure gradient (APG). The flow expends its kinetic energy to overcome the APG as it decelerates, and eventually separates from the surface. It is important to reduce flow separation to improve and maintain aerodynamic efficiency, and the approach of interest is to energize the flow to help overcome the APG. Passive VGs aid in reducing flow separation by entraining the turbulent kinetic energy (TKE) from the free-stream flow to the near wall region. Prior research in passive flow control reveals that the effectiveness of a VG in controlling separation depends on multiple factors which include, the size of the VG relative to the boundary layer thickness, spacing between adjacent VGs, and position of the VG with respect to the line of separation. While recent advances in numerical methods and computational resources have brought more complex flows under our computational grasp, resolving all the length and time scales for a large portion of real-world flows is still unfeasible. Large Eddy Simulations (LES) provide a promising alternative and is our tool for investigation in this study.
An optimal deployment of cubes to control boundary layer separation requires a thorough understanding of the TKE entrainment and distribution in the wake of the cubes. The dependence of these quantities on the cube to height to boundary layer thickness ratio and spacing between adjacent cubes is poorly understood. Therefore, the objectives of this work are to perform LES of flow over wall-mounted cubes in a TBL to understand the effect of: (i) cube height to boundary layer thickness ratio, and (ii) inter-cube spacing on the near-wake characteristics in general, and TKE distribution in particular. To achieve these objectives, we validate an existing approach to simulate a spatially evolving turbulent boundary layer (SETBL), and propose a novel method using machine learning for the same purpose, with the aim of reducing computation time without any significant modification to the numerical framework. For a single cube placed in SETBL on a flat plate we discover that the TKE per unit area decays as a power law in the near-wake, and the power law exponent increases in a non-linear manner with increasing cube height. LES of flow over an array of cubes in SETBL reveals amplification of large scale coherent structures in the outer region of the TBL which are characterized by increasing TKE. We believe the ejection of low momentum fluid in the region in between adjacent cubes is responsible for this amplification. Our findings have direct applications in reducing aerodynamic drag on automobiles, aircrafts and improving turbine efficiency, which in turn can help us reduce greenhouse gas emissions.