Problem with pseudo shock simulation!

Sridhar thiru · September 5, 2012, 06:38

Hi Guys!! am doing simulation on constant area straight ducts which mimics supersonic inlet isolators, i would like the study the pseudo shock phenomenon that occurs due to shock wave boundary layer interactions.

My problem domain is a cuboid with dimension 80x80x960 all in mm..
owing to computation i reduced the cross section dimension to 40x40 mm, with two long faces of cuboid set to symmetry. n length remained same.

Grid : cuboid with above dimensions;
structured grid with grid fined near the corner region,
1.5 million hexahedral cells, cell squish is around 0.2

Solver settings:
Density based, implicit, absolute velocity formulation.
Material, Air as ideal gas, viscosity :sutherlands formulation.
Turbulence model: standard K - omega model with low re number effect & shear flow corrections enabled,

Boundary conditions are as follows,

Pressure inlet :
momentum- tot pressure 196000pa
supersonic int gauge pressure - 25050 pa
Turbulence parameters
Hydraulic diameter 0.08m
intensity 5 %
Total Temperature : 300k

pressure outlet :
momentum-gauge pressure - 25050 pa
Turbulence parameters
Hydraulic diameter 0.08m
intensity 5 %
Total temperature : 300k

wall : no slip & stationary, heat flux set to zero (adiabatic!)

symmetry : default symmetry

Solver is a STEADY SOLVER ( just to obtain initial conditions for UNSTEADY SOLVER)

discretization:
flux : AUSM
flow : third order MUSCL
Turbulence scalar equations : QUICK
CFL set to default value of 5.

wall y+, i calculated using NASA viscous grid spacing calculator i fixed y+ of 30 which throws cell height of 1e-5m!! Re.no is 3.8*e6 with reference length as hydraulic diameter..

The problem is my solution is not converging beyond 1e-5, residuals fluctuate around 1e-5.. i set the convergence criteria to 1e-6.!!

Is it enough to switch to unsteady solver???
If that the case what value of time step i provide?? also max iterations per time step..I'm stuck with this, Help me..

Also when check with the contour, fluent shows the max wall y+ value of 19 near the wall though i set the cell height for y+ 30.

Thanks in advance.

Shamoon Jamshed · December 22, 2015, 08:42

Hi,

Although seen your message very late, but may be its still helpful. It appears that you are using Fluent (although not explicitly mentioned). The modelling parameters seem ok. The oscillations in residual is not a big deal. This usually comes because the shock at each iteration is changing its position (very very little bit) which does not affect the overall solution. Try to put a monitor in or very near to the shock region and monitor (e.g Pressure or velocity) across. Then you will see some periodic pattern in the monitor output, and if it shows then our solution is fine. and you may , if you wish continue to run unsteady if you want to see some turbulence features. However, the steady state solution is converged. Secondly, monitor in fluxes, that the difference b/w mass flow inlet and outlet is zero, so that the mass is conserved and the continuity satisfies.
Also check the contours of pressure or velocity after (e.g 50) iterations.

September 5, 2012, 06:38	Problem with pseudo shock simulation!	#1
Sridhar thiru New Member Sridhar Join Date: Aug 2012 Posts: 2 Rep Power: 0	Hi Guys!! am doing simulation on constant area straight ducts which mimics supersonic inlet isolators, i would like the study the pseudo shock phenomenon that occurs due to shock wave boundary layer interactions. My problem domain is a cuboid with dimension 80x80x960 all in mm.. owing to computation i reduced the cross section dimension to 40x40 mm, with two long faces of cuboid set to symmetry. n length remained same. Grid : cuboid with above dimensions; structured grid with grid fined near the corner region, 1.5 million hexahedral cells, cell squish is around 0.2 Solver settings: Density based, implicit, absolute velocity formulation. Material, Air as ideal gas, viscosity :sutherlands formulation. Turbulence model: standard K - omega model with low re number effect & shear flow corrections enabled, Boundary conditions are as follows, Pressure inlet : momentum- tot pressure 196000pa supersonic int gauge pressure - 25050 pa Turbulence parameters Hydraulic diameter 0.08m intensity 5 % Total Temperature : 300k pressure outlet : momentum-gauge pressure - 25050 pa Turbulence parameters Hydraulic diameter 0.08m intensity 5 % Total temperature : 300k wall : no slip & stationary, heat flux set to zero (adiabatic!) symmetry : default symmetry Solver is a STEADY SOLVER ( just to obtain initial conditions for UNSTEADY SOLVER) discretization: flux : AUSM flow : third order MUSCL Turbulence scalar equations : QUICK CFL set to default value of 5. wall y+, i calculated using NASA viscous grid spacing calculator i fixed y+ of 30 which throws cell height of 1e-5m!! Re.no is 3.8*e6 with reference length as hydraulic diameter.. The problem is my solution is not converging beyond 1e-5, residuals fluctuate around 1e-5.. i set the convergence criteria to 1e-6.!! Is it enough to switch to unsteady solver??? If that the case what value of time step i provide?? also max iterations per time step..I'm stuck with this, Help me.. Also when check with the contour, fluent shows the max wall y+ value of 19 near the wall though i set the cell height for y+ 30. Thanks in advance.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
the shock tube problem and its artifacts with sonicFoam	emjay	OpenFOAM Running, Solving & CFD	0	December 16, 2011 08:58
Turbine: Transient Simulation Problem	John C.	CFX	7	December 5, 2011 09:31
[ask] shock absorber simulation	dik	Main CFD Forum	1	December 17, 2009 02:32
CFX converge problem caused by shock waves	littlelz	CFX	3	August 17, 2009 10:35
tracer simulation for transient problem.	John	CFX	0	August 12, 2005 11:37

December 22, 2015, 08:42		#2
Shamoon Jamshed Senior Member Shamoon Jamshed Join Date: Apr 2009 Location: Karachi Posts: 381 Rep Power: 18	Hi, Although seen your message very late, but may be its still helpful. It appears that you are using Fluent (although not explicitly mentioned). The modelling parameters seem ok. The oscillations in residual is not a big deal. This usually comes because the shock at each iteration is changing its position (very very little bit) which does not affect the overall solution. Try to put a monitor in or very near to the shock region and monitor (e.g Pressure or velocity) across. Then you will see some periodic pattern in the monitor output, and if it shows then our solution is fine. and you may , if you wish continue to run unsteady if you want to see some turbulence features. However, the steady state solution is converged. Secondly, monitor in fluxes, that the difference b/w mass flow inlet and outlet is zero, so that the mass is conserved and the continuity satisfies. Also check the contours of pressure or velocity after (e.g 50) iterations.