Date of Award

May 2023

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Engineering

First Advisor

Ryoichi Amano

Committee Members

Ilya Avdeev, Istvan Lauko, Wilkistar Otieno, John Reisel

Keywords

Computational Fluid Dynamics, Crossflow Mitigation, Heat Transfer, Jet Impingement Cooling, Numerical Simulations

Abstract

Jet impingement is a cooling method used to efficiently cool gas turbine blades and other engine components. In jet impingement cooling, fluid jets travel perpendicular to the hot surface through nozzles and impinge on the surface with high velocity to dissipate the thermal load. Internal cooling of gas turbine blades is commonly performed with impingement cooling and serpentine channels. Besides gas turbine blades, the other turbine components such as turbine guide vanes, rotor disks, and combustor walls can be cooled using high-momentum air jets. Additionally, jet impingement is not only limited to gas turbine cooling, since it is also applicable for electronic-chip cooling, hot metal sheets cooling, paper and fabric drying, as well as many other applications.As providing compressed cooling air can be quite expensive, this study is focused on optimizing the cooling air characteristics to maximize heat transfer and enhance cooling efficiency. The study incorporates three scopes; the first highlights the main geometrical and flow aspects that affect the cooling performance and can potentially influence the design of the cooling circuits. These geometrical and flow factors include Reynolds number (??), jet-to-target spacing -also known as stand-off distance (?/?)-, jet array configuration (in-line and staggered), and jet-to-jet spacing (?/?) with an intermediate crossflow condition (double exit case), where the spent air has a bidirectional exit. A range of Reynolds number (??) of 2,500 – 12,500 was investigated for air jets issued from a 5 × 5 square nozzle array and correlated with the Nusselt number (??) for laminar and turbulent flow regimes. It was found that the increase in Reynolds number (??) resulted in a higher stagnation Nusselt number (??), where it peaked underneath the middle row of jets for the intermediate crossflow scheme (bi-directional exit). A detailed comparison of different stand-off distances (?/?) of 3, 5, and 7.5 for each flow case showed a reduction in both line-averaged and area-averaged Nusselt number (??) as the stand-off distance (?/?) increased. The increased stand-off distance results in a higher entrainment of the surrounding air that causes the loss of jets’ momentum. Jet configuration was also explored by comparing in-line and staggered nozzle arrays operating at the same conditions of Reynolds number (???) and stand-off distance (?/?). It was concluded that the capability of achieving higher heat transfer rates of the in-line nozzle configuration is higher than the staggered one, especially at shorter stand-off distances (?/?). In in-line arrays, the jet rows tend to protect each other from the crossflows interacting with the main jet streams and weakening them, due to the shorter distance given to air crossflow to accelerate before striking the next row of jets. However, in the staggered arrangement, the influence of the crossflow is higher as the crossflow accelerates further and directly impacts the jets. Additionally, jet interactions were also investigated through jet-to-jet spacing (?/?) as manipulating the spacing between the jets can change how each two neighboring jet streams interact with one another. It was concluded that as the open area ratio (??) decreases, the heat transfer performance decreases due to the enlarged area that the nozzles are required to cover. The second scope of this study deliberates the jet-to-jet interactions and the accumulation of the crossflow as the spent air migrates downstream before exiting the fluid domain. The crossflow accumulation usually causes a decay in the Nusselt number and, hence, reduces the overall cooling performance. This study investigated several crossflow mitigation techniques for a maximized crossflow condition (single exit case), where the spent air has a unidirectional exit. Firstly, by attaching small and easy-to-manufactured crossflow diverters to protect jet streams from upstream crossflows and, therefore, enhance heat transfer performance. The investigations comprised three diverter shapes of cylindrical, rectangular, and ribbed type with heights of 25%, 50%, and 75%, as a percentage of jet-to-target (?/?) spacing denoted as Quarter-Length (QL), Half-Length (HL), and Three-Quarter-Length (TQL), respectively. Secondly, by extending the jets, where the jet shells were extended in the flow field, where the jets were either of a fixed height of 1.5D (EXTF) or with a variable height of 0 - 2D (EXTV). Results considered the increase in Nusselt number (??) along with the associated pressure loss and showed a net enhancement of up to 5.2%, 3.1%, and 3.8% for cylindrical, rectangular, and ribbed-type diverters, respectively. Additionally, the EXTF showed the highest net heat transfer enhancement among all other techniques. It offered an average of 13.4% net enhancement while the EXTV offered 10.4%. Thirdly, by manipulating the jet diameters in a 5 × 5 in-line array where the jet diameters were varied according to three scenarios of 5%, 10%, and 20% increase towards the spent air exit. The variable jet diameter showed a minimal enhancement of Nusselt number (??) not exceeding 2.09 % for the 5% increase in the jet diameter, while it adversely affected the heat transfer for other scenarios. The third scope is to experimentally study the effect of rotation on the temperature surface profile under three rotational speeds (rpm) and three different Reynolds numbers (??) in the turbulent flow regime and under non-uniform heat flux conditions. It was concluded that as the rotational speed increased, the jet diversion was increased due to the increased tangential force component. This diversion in jets was minimized as the jets’ momentum increased (increase in jet Reynolds number (??)). Overall, most of the study scopes involved developing numerical CFD codes using StarCCM+ software to validate the experimental results so that the same model can be used to numerically investigate other geometries and aspects. Improving a reliable CFD model with an acceptable level of accuracy and simulation convergence time led to eliminating the need to conduct expensive and complicated experiments. The k-ε turbulence model was used due to its robustness and fast convergence time. As this study explored a wide range of thermofluidic parametric analyses of jet impingement, resolving turbulence scales was not a priority, and the steady state assumption while modeling turbulence is suitable for the application. Additionally, post-processing techniques for analyzing thermal images were heavily implemented to handle, process, and present the results.

Download

Available for download on Monday, May 26, 2025

Share

COinS