Simulation of a complete shock tunnel using parallel computer codes

Goozee, Richard J. (2003). Simulation of a complete shock tunnel using parallel computer codes PhD Thesis, School of Engineering, The University of Queensland.

Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
THE17470.pdf Full text application/pdf 54.04MB 0
Author Goozee, Richard J.
Thesis Title Simulation of a complete shock tunnel using parallel computer codes
School, Centre or Institute School of Engineering
Institution The University of Queensland
Publication date 2003
Thesis type PhD Thesis
Supervisor Dr Peter Jacobs
Total pages 349
Collection year 2003
Language eng
Subjects L
290299 Aerospace Engineering not elsewhere classified
690302 Space transport
Formatted abstract
Impulse facilities provide a unique capability in being able to reproduce aspects of the hypersonic flight environment in the laboratory. The usefulness of these facilities is limited by non-ideal aspects of their operation and by significant gaps in our understanding of these flows. In this thesis, simulations have been performed of a reflected shock tunnel facility, operated at the University of Queensland, with the aim of providing a better understanding of the flow through these facilities. In particular, the analysis of the simulations focuses on the premature contamination of the test flow with driver gas and the generation of the high levels of noise experienced in the test flow. The substantial computational effort required by these calculations has, in the past, meant that only parts of a facility have be modelled in any one simulation. The assumptions associated with modelling only part of the facility has meant that these simulations have not been successful in predicting either driver gas contamination or noise levels.

The multi-block Computational Fluid Dynamics (CFD) code MB_CNS, which is based on a finite-volume formulation of the compressible Navier-Stokes equations was used in the calculations. As an alternative numerical technique, Smoothed Particle Hydrodynamics (SPH) was investigated for advantages that it may provide in modelling flows involving gaseous interfaces. It was determined that this technique is limited in its applicability to shock tunnel flows due to problems including the treatment of solid boundary conditions.

The computational requirements of simulating a complete shock tunnel have been approached with the use of large parallel supercomputers. Both MB_CNS and the SPH code were parallelised using the shared memory approach of OpenMP and the distributed memory approach of MPI. The performance of the parallel codes are examined on various computers, including the APAC National Facility and the QPSF Facility, located at the University of Queensland.

The shock wave induced deformation of bubbles, of both a light and heavy gas, is examined as a test case for the shock tunnel modelling. This case has experimental photographs that can be compared directly with the simulations. These simulations demonstrate that MB_CNS can accurately model the shock induced instability and deformation of interfaces between different gases.

The simulations presented here assume an axisymmetric flow and mesh the complete facility, from the driver section to the dump-tank. By doing this, the current simulations eliminate many of the assumptions previously made in shock tunnel simulations. The simulations incorporate an iris-based model of the rupture mechanics of the primary diaphragm, an ideal secondary diaphragm and account for turbulence in the shock tube boundary layer with the Baldwin-Lomax eddy viscosity model. Supporting these simulations are three sets of experiments, with different operating characteristics, that were conducted by Dr. D. Buttsworth in the Drummond Tunnel facility. The results from these experiments are used as validation for the simulations.

Through the shock tunnel simulations, a better understanding of the mechanisms leading to driver gas contamination has been developed. It has been shown that the contamination is driven by a complex interaction between the reflected shock and the incident gases, which result in the generation of vorticity in the driver gas. In the tailored case that was simulated, a vortex ring was shown to form at the head of the driver gas, which moved along the centreline of the shock tube into the test flow resulting in the observed contamination. Due to the complex, transient nature of this process, it could not be predicted using simplified models, or even numerical simulations using only a part of a facility. The instability resulting from the interaction between the reflected shock and the contact surface is driven by the same mechanisms that are studied in the simulations of the shock wave interaction with the bubbles.

It is shown that the simulations performed in this thesis can provide an estimate of the noise levels experienced in the test flows produced by shock tunnels. This estimation is possible since the simulations include some of the noise producing mechanisms occurring throughout the facility. The generation of this noise is demonstrated, along with its propagation into the test flow. The noise levels measured in the simulated test flow were in agreement with noise levels measured in the test flow of the Drummond Tunnel.
Keyword Hypersonic wind tunnels
Additional Notes Missing page 112, 113, 136, 137, 144, 145, 148, 149, 294, 295 in the original thesis.

Document type: Thesis
Collection: UQ Theses (RHD) - UQ staff and students only
Citation counts: Google Scholar Search Google Scholar
Created: Fri, 24 Aug 2007, 18:11:21 EST