Effective boundary conditions for transpiration cooling applications

• Effektive Randbedingungen für Transpirationskühlungsanwendungen

König, Valentina; Müller, Siegfried (Thesis advisor); Herty, Michael (Thesis advisor)

Aachen : RWTH Aachen University (2021, 2022)
Dissertation / PhD Thesis

Dissertation, RWTH Aachen University, 2021

Abstract

Using transpiration cooling with carbon/carbon (C/C) thrust chamberliner is identified as a new innovative cooling concept that can lead to improvement in advanced rocket engines. In addition to experiments, computational fluid dynamics simulations offer an efficient and low costpossibility to investigate the physical phenomena of transpiration cooling. In the present work an effective model is developed that simulates transpiration cooling taking microscale effects at the interface between ahot gas flow and a porous medium flow into account without resolving the microscale pores. The derivation of our general strategy is based on upscaling and consists of three models: the multiscale model, the zeroth-order model and the effective model, where the latter two models operate on the macroscale. Here the multiscale model captures the local injection of a coolant through a large number of pore size injection channels. It is set up to derive appropriate cell problems on the microscale and to validate the effective model. For the latter effective boundary conditions are developed using an upscaling approach. To validate the effective model numerical computations are presented. Furthermore, the influence of the microscale characteristics on the heat transport in turbulent flow over a porous material is investigated. All computations are based on wind tunnel experiments performed at the ITLR Stuttgart with a porous C/C sample produced at the DLR Stuttgart. For the injection rate F = 0.1% the numerical solutions of the three modelsare compared to each other in terms of temperature distribution, wall shear stress, wall heat flux and cooling efficiency. Numerical computations show that the predicted cooling efficiency is reduced when using a local injection (multiscale) in comparison to a uniform injection (zerothorder). This effect is reflected in the effective computation. Thus, the effective model provides a more accurate approximation than the zeroth order solution. Furthermore, the effective model is significantly more efficient compared to a fully resolved multiscale computation. This is confirmed by comparing the amount of grid cells and computational times.