DSMC Simulations of Rarefied Flow Over A Square Cylinder
DSMC Simulations of Rarefied Flow Over A Square Cylinder
13:21
iccms14-paperX
The aero-thermal analysis of different shaped geometries in rarefied hypersonic flow fields
are important and find applications in the design of devices to measure the hypersonic flow
properties, elements of satellite systems, re-entry conditions etc. In this work, rarefied gas
flow of Argon around a square cylinder is simulated using Direct Simulation Monte Carlo
(DSMC) method. The simulation is carried out for different free-stream Knudsen number
such as 0.002, 0.01, 0.05 & 0.25. The effect of angle of attack of the entry flow is also
analyzed. The results such as variation of pressure, temperature, local Knudsen number and
local Mach number in the flow domain, shock layer - boundary layer interactions and its effects
on rarefaction, drag force, heat and location of the peak heat flux over the surface, velocity slip
and temperature jump over the surface of the square cylinder are presented and discussed in
this work.
Keywords: Rarefied flow; hypersonic flow; DSMC simulation & square cylinder.
Introduction
Characteristics of flow behavior in the rarefied gas regime is very important in the design of
vehicles operating at higher altitudes. Vehicles at high altitudes operate at hypersonic speed
and experiences a variety of complex flow features. The concept of continuum is no longer
valid in rarefied flows. The ratio of mean free path of the gas molecule and characteristic
length is called Knudsen number, is the non-dimensional parameter defining the degree of
rarefaction in a flow. The solution of Navier-Stokes equation becomes invalid for Knudsen
number, Kn > 0.1. Boltzmann equation is the equation that is valid for all flow regimes. But
Boltzmann equation is too difficult to solve and this difficulty has led to the development of a
computationally easy technique called Direct Simulation Monte Carlo (DSMC) Method [Bird
(1994)]. It is a statistical method which emulate the physical processes modeled by Boltzmann
equation.
DSMC method is a particle based simulation in which a simulated molecule represents a
much larger ensemble of molecules tracked through space and time. The movement of
particle and collision between particles are computationally decoupled. The movement of
each particle is linear in nature as prescribed by its velocity vector and time-step increment.
1
13:21
iccms14-paperX
The reorientation in the particle trajectory occurs as a result of collision with the wall and the
intermolecular collisions. The particle motion and surface collision are deterministic whereas
intermolecular collisions are probabilistic in nature. The collision rate is set same as that
of kinetic theory of gas and probability of collision between collision pairs is found out by
acceptance-rejection method. The collision is modelled using available phenomenological
models. The macroscopic properties are found out by taking the ensemble average of the
properties of particles.
DSMC Solver - dsmcFoam
OpenFOAM [Jasak et al. (2007)] is an opensource fluid flow solver based on object oriented
C++ modules and includes variety of solvers which are capable of handling most of the physical problems. dsmcFoam [Scanlon T.J. et al. (2010)] (DSMC solver available in OpenFOAM)
has the following capabilities viz. steady and unsteady simulations by proceeding through
small time steps, capability to simulate arbitrary 2D and 3D geometries, ability to define
symmetry plane and cyclic boundaries, arbitrary number of gas species can be included in the
simulation, free-stream flows over the obstacles. The Larsen-Borgnakke model and variable
hard sphere (VHS) model for DSMC collision simulation are available in the module.
Validation case - Hypersonic flow over a wedge
The work presented by Liu et al. (2012) for a hypersonic flow over wedge of angle 40o was
simulated using the dsmcFoam as a validation case. The computational domain and boundary
conditions of the problem for validation case is shown in figure 1. The inlet Mach number
of 10 is simulated with an angle of attack of 10o as shown. The free-stream temperature is
taken as 200 K. The wedge surface is yaken as a thermal wall maintained at a temperature of
300 K. The free-stream Knudsen (Kn) number is computed by taking wedge base length as
the characteristic length, which is 0.05. The properties of Argon gas molecules taken for the
computations and the physical boundary condition and other conditions are given in the table
1 and table 2 respectively. The time step of computations is 1 106 s. The equivalent number
of particle is taken as 5 1014 for a number density of 1.373 1020 .
Table 1.
wedge.
The molecular properties taken for Ar gas for the simulation of hypersonic flow over a
Tref (K)
m(1010 )(kg)
1000
0.734
3.595
66.3
The predicted temperature and density contours and those presented by Liu et al. (2012)
within the domain for a free-stream Mach number of M = 10 and angle of attack, 10o are
shown in in figures 2 and 3. A comparison of normalized surface pressure distributions over
the inclined walls and comparison with Liu et al. (2012) is shown in fig. 4. A good agreement
between the two results can be seen.
2
13:21
iccms14-paperX
V (m/s)
T (K)
Tb (K)
10
2635
200
300
1.373 1020
Fig. 1.
The computational domain for the analysis of hypersonic flow over wedge
Fig. 2.
(b) dsmcFoam
13:21
iccms14-paperX
Fig. 3.
Fig. 4.
(b) dsmcFoam
13:21
iccms14-paperX
Fig. 5.
The computational domain for the analysis of hypersonic flow over square cylinder
Fig. 6 shows the distribution of temperature and velocity along the stagnation line. The shock
stand-off distance and shock thickness is higher in the case with Kn=0.25. An increase of 156
K is obtained in the peak temperature for low values of Knudsen number. At the downstream
of square cylinder values of temperature is less due to higher rarefied region. The surface
parameters such as pressure, shear stress and heat flux are obtained from the simulation. The
total drag and heat on the wall surface for different cases are compared. The effect of Knudsen
number and angle of attack on slip velocity and temperature jump are also obtained from the
analysis. Different Kn - shock stand-off distance.- Graph
(a) Kn=0.002
Fig. 6.
(b) Kn=0.01
(c) Kn=0.05
(d) Kn=0.25
13:21
iccms14-paperX
(a) Kn=0.002
(b) Kn=0.01
Fig. 7.
(c) Kn=0.05
(d) Kn=0.25
Table 3.
Overall performance parameters for the simulation of hypersonic flow over a square cylinder.
Kn
Pressure
Drag(N)
Total Heat
per unit area
Peak pressure(Pa)
Peak heat
(W/m2 )
0
0
0
0
0.002
0.01
0.05
0.25
9.72
1.89
0.39
0.082
2313
396
115
39
875
173
35
7.09
262000
39700
11600
3840
15
15
15
15
0.002
0.01
0.05
0.25
10.2
1.98
0.4
0.085
2472
424.74
121.47
41.27
897
173
34.7
7.21
334000
53400
15200
4890
30
30
30
30
0.002
0.01
0.05
0.25
11.65
2.3
0.46
0.094
2927.4
505.2
142.4
47.73
1020
186
37.3
7.74
412000
71900
20000
6060
45
45
45
45
0.002
0.01
0.05
0.25
13.55
2.67
0.53
0.11
3328.9
580.86
162.53
353.69
854
169
33.5
6.3
34200
70800
20300
5560
13:21
iccms14-paperX
Fig. 8. Contours of Mach number, pressure (Pa) and temperature (K) for Kn = 0.01 and =
0o , 15o , 30o &45o
13:21
iccms14-paperX
Conclusions
The DSMC simulation is carried out for rarefied hypersonic flow of Argon gas over a square
cylinder. The effect of Knudsen number, angle of attack and associated flow characteristics are
discussed. The variation of parameters in the flow domain are presented in the form of colour
contours. The distribution of slip velocity, temperature jump, pressure, shear stress, and heat
flux over the wall boundary is also reported.
References
1. Bird G.A., (1994). Molecular Gas Dynamics and the Direct Simulation of Gas Flows, Clarendon
Press, Oxford Science Publications.
2. Scanlon T.J., Roohi E., White C., Darbandi M., Reese J.M., (2010), An open source, parallel DSMC
code for rarefied gas flows in arbitrary geometries, Computers and Fluids, Volume 39-10, 2078-2089.
3. Lofthouse A.J., (2008), Non equilibrium hypersonic aerothermodynamics using Direct simulation
Monte Carlo and Navier-stokes models, Ph.D. Thesis, University of Michigan.
4. Bird G.A., (2006), The DS3V Program Users Guide, Version. 2.5, GAB Consulting Private Ltd,
Sydney.
5. Jasak H, Aleksandar J., Zeljko T, (2007), OpenFOAM: A C++ library for complex physics simulations International workshop on coupled methods in numerical dynamics, Vol. 1000, 1-20.
6. Liu, H., Cai, C., & Zou, C. (2012), An object oriented serial implementation of a DSMC simulation
package, Computers & Fluids, 57, 66-75.