of 17
Measurements of hypersonic double cone flows with shock
wave/boundary layer interactions in the X3 expansion tunnel
Aaron Kennedy
and Rowland Penty-Geraets
University of Oxford, Oxford, Oxfordshire, OX2 0ES, UK
Christopher M. James
, Matthew Thompson
§
and Richard G. Morgan
The University of Queensland, Brisbane, Queensland 4072, Australia
Joanna M. Austin
†
California Institute of Technology (GALCIT), Pasadena, California 91125
Fabian Zander
∗∗
University of Southern Queensland, Toowoomba, Queensland 4350, Australia
Matthew McGilvray
††
University of Oxford, Oxford, Oxfordshire, OX2 0ES, UK
This paper presents the results from a series of hypersonic double cone experiments,
conducted in the X3 expansion tunnel at the University of Queensland (UQ). The model was
previously used and produced at California Institute of Technology (Caltech) and is a scaled
model of the double cone model used by the Calspan-University of Buffalo Research Center
(CUBRC). The data presented is a full transient set of surface heat flux, and surface pressure
measurements along the axial length of the model as well as a full suite of tunnel data. These
aim to help characterize the flow over the geometry by providing surface conditions, an accurate
location of the boundary layer separation, the flow reattachment points and the general structure
of the shockwave boundary layer interaction. The geometry of the X3 facility is also provided to
give a more complete picture of the test conditions, an important consideration for numerical
models. The outcome of this paper is to provide the most detailed experimental data to date
and improve validation attempts of existing non-equilibrium thermochemistry codes.
I. Nomenclature
= Free stream speed of sound m.s
1
0
= Total enthalpy J.kg
1
= Wall enthalpy J.kg
1
= Free stream static enthalpy J.kg
1
= Free stream Mach number
푓푙
= Number of flow lengths of gas untill steady Pitot conditions
퐶푇
= Compression tube pressure Pa
퐷푇
= Dump tank/acceleration tube pressure Pa
푃푖푡표푡
= Pitot pressure Pa
푃푖푡표푡
푒푥
= Average experimental Pitot pressure Pa
D.Phil Candidate, Department of Engineering Science, Oxford Thermofluids Institute.
Research Associate, Department of Engineering Science, Oxford Thermofluids Institute.
ARC DECRA Fellow, Centre for Hypersonics, School of Mechanical and Mining Engineering, AIAA Senior Member.
§
PhD Candidate, Centre for Hypersonics, School of Mechanical and Mining Engineering.
Professor, Centre for Hypersonics, School of Mechanical and Mining Engineering. Associate Fellow AIAA.
†
Professor of Aerospace, Graduate Aerospace Laboratories, 1200 E. California Blvd, MC 105-50, Associate Fellow AIAA.
∗∗
Associate Professor, Institute of Advanced Engineering and Space Sciences. Member AIAA.
††
Associate Professor, Department of Engineering Science, Oxford Thermofluids Institute.
1
푅푒푠
= Reservoir pressure Pa
푠푡푎푡푖푐
= Free stream static pressure Pa
푆푇
= Shock tube pressure Pa
0
= Total pressure Pa
¤
= Wall heat flux W.m
2
푅푒
= Reynolds number behind the primary conical shock wave
푅푒
= Unit Reynolds number behind the primary conical shock wave m
1
푅푒
= Free stream unit Reynolds number m
1
푅푒
= Free stream Reynolds number
푆푡
= Stanton number calculated with post primary conical shock wave properties
푆푡
= Stanton number calculated with free stream properties
= Free stream temperature K
1
= shock tube shock wave speed m.s
1
2
= Acceleration tube shock wave speed m.s
1
= Flow velocity behind the primary conical shock wave m.s
1
= Free stream velocity m.s
1
= Density behind the primary conical shock wave kg.m
3
= Free stream density kg.m
3
II. Introduction
The double cone geometry has been used in various test cases for several years now, ranging from use exploring
the effects of indented ablated nose tips [
1
], exploring real gas effects and more. The model used for the experiments
presented in this paper was derived from the model first used by Holden et al. [
1
] as a validation test case for numerical
codes. However, the flow a phenomenon created by this geometry is also applicable to the real world in locations where
compression corners would exist such as control surfaces. While the general shock structure is analogous to a ‘Type VI’
shock interference case [
2
], the compression corner causes flow separation at high enthalpies in the corner region which
produces both separation and reattachment shocks making a complicated flow structure. This produces larger variations
in the heat flux and surface pressure along the surface of the geometry with peak values occurring in the region of flow
reattachment. Any future attempts at designing hypersonic vehicles would greatly benefit from the full understanding
and capability to predict flow properties on these features, thus there is a real-world motive for this test case considering
the geometry alone.
Fig. 1 Double Cone model during test shot in the T6 tunnel at the University of Oxford.
There have since been multiple experimental sets by the group at Calspan-University of Buffalo Research Center
(CUBRC) [
3
,
4
], which found cases run in nitrogen test gas proved successful in validation attempts by numerical
modellers. However, in air environments, predictions showed poor agreement in higher enthalpies particularly for the
prediction of the size of the separation region. Subsequent tests in pure oxygen environments showed similar results to
the air cases with spectroscopy showing poor agreement with the level of nonequilibrium thermochemistry predicted in
the free stream [
4
]. This was addressed by switching the experimental methods from reflected shock tunnel (LENS I) to
2
an expansion tube (LENS XX) where the level on nonequilibrium in the freestream would be less due to the principals
of operation of expansion tubes. This led to the latest data set to be released by the group at CUBRC presented in [4].
Further prediction attempts completed [
5
8
] still show difficulties in predicting the flow separation lengths with
attempts of uncertainty analysis [
9
] also completed to try and reach a conclusion. There are further datasets that were
added to the mix such as the work conducted at Caltech T5 by Knisley and Austin [
10
], and the experiments at the
Arnold Engineering Development Center (AEDC) tunnel no. 9 [
11
]. The tunnel no. 9 experiments are of interest as
they describe unstable flow behaviour throughout the test time.
With the preceding work showing difficulties in the accurate prediction of flow properties, it shows further work is
necessary to fully understand the problem. The aim of the work presented in this paper is to provide a complete transient
dataset for the research community to use to improve the code validation attempts. One of the largest improvements on
previous data sets is the addition of details of the X3 facility, shock speed history and the temporal variation of the
measurements throughout the test time. It is also believed by the authors, that significant uncertainties surrounding the
flow processes inside expansion tube facilities and the resultant flow produced. A secondary objective of this paper is to
present initial data to be used to develop further understanding on the fundamentals of the facilities going forward.
III. Experimental setup
A. Experimental Test Facility
Fig. 2 Schematic of X3 in expansion tube mode from Stennett et al. [12]
Figure 2 shows the layout of X3 in the expansion tunnel mode for this test campaign. The development of X3 by
Prof. R. G. Morgan began in 1994 with the first commissioning shots fired in 2001. For a detailed summary of the
history X3 the reader is directed to Gildfind et al. [
13
]. The compression tube was separated from the shock tubes via a
stainless steel primary diaphragm at the ‘primary diaphragm station’ of a thickness and score depth dependent on the
desired rupture pressure. Both shock tubes 1 and 2 were used as one large shock tube filled with zero air as the primary
test gas. Finally, the acceleration tube and dump tank formed a large continuous volume that was evacuated via vacuum
pumps to a desired pressure. The acceleration tube and shock tube separated by a 13
m Mylar secondary diaphragm
located at the ‘tertiary diaphragm station’ shown on Figure 2. The Mylar thickness was chosen to be as thin a practically
possible to mitigate the chance of model surface damage post rupture. An example of this effect can be seen in the work
by James et. al [14].
In 2015, the X3 facility began development to incorporate a reflected shock tunnel (RST) mode of operation within
its capability skill set to become a dual mode facility. This development saw modification to X3 including the re-boring
of the second shock tube (st2) to 200 mm internal diameter, equal to shock tube 1. The acceleration tube has remained at
the original 182.6 mm diameter meaning there is an area reduction from 200 mm to 182.6 mm at the interface between
the shock tube and acceleration tube. Prior to the re-boring, there was an area increase at this same location. The effect
of the area reduction on the test flow is unknown at this stage but a ramp section has been installed at this interface to
avoid an abrupt change and therefore mitigate any effect that could be present. However, it is likely the presence of this
area reduction serves to increase the uncertainty in the flow conditions by a significant amount. During this period of
modification, the tunnel was relocated to a new site owned by the Department of Defence, Science and Technology
(DST) in Brisbane, Australia giving it significantly more space to operate.
The current facility geometry including the area reduction at the secondary diaphragm used for this test campaign is
presented in Figure 3. Due to the aspect ratio of the plot, the tunnel profile presented in Figure 3 exaggerates the area
reduction to look almost like a forward-facing step. However, in reality the reduction occurs via a slope of approximately
6
.
3
.
3