Javascript must be enabled for the correct page display

An efficient time integration method for extra-large eddy simulations

Scheijbeler, M. (2005) An efficient time integration method for extra-large eddy simulations. Master's Thesis / Essay, Mathematics.

[img]
Preview
Text
Marieke_Scheijbeler_doctoraal.pdf - Published Version

Download (3MB) | Preview

Abstract

The calculation of the dynamic loads on space launchers is important for the design and optimisation of space launchers. With aerodynamic calculations these time-dependent aerodynamic loads could be calculated using Extra-Large Eddy Simulations (X-LES). X-LES simulations solve more details of the flow physics than Reynolds-averaged Navier–Stokes (RANS), without the costs of a full Large Eddy Simulation (LES). Nonetheless, these time-accurate simulations cost a lot of computational time, therefore an efficient time integration method has been designed. It takes thousands of time steps to get statistically converged data for X-LES. For these time steps, implicit time integration, based on dual time stepping method is currently used. In the LES region the physical (accuracy) time step is of the same order as the numerical (stability) time step. Therefore it is more efficient to use an explicit time integration method, with a few relaxations in the LES region. The grid cells in the RANS region, especially the boundary layer, are much smaller than in the LES region, hence in the RANS region an explicit time integration method is not efficient. In the RANS region the implicit time integration method based on the dual time stepping will be used, but in the LES region the time integration method will be replaced with an explicit method. The coupling between the time integration method on the interface of the RANS and the LES region will be designed. The coupling that has been designed is second order accurate, stable and conservative. These aspects are analysed in this document. The coupling is implemented in a multi-block Euler/Navier–Stokes flow solver. Test cases show that if the implicit and explicit region are of the same size, then the speedup is two with respect to a fully implicit calculation.

Item Type: Thesis (Master's Thesis / Essay)
Degree programme: Mathematics
Thesis type: Master's Thesis / Essay
Language: English
Date Deposited: 15 Feb 2018 07:28
Last Modified: 15 Feb 2018 07:28
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/8369

Actions (login required)

View Item View Item