Channel Flow using CONVERGE CFD

Report: Channel Flow using CONVERGE CFD

Case description:

Fig 1

• Just to test that all software is installed and running properly.
• This is simple 2D simulation of a channel flow with the boundaries as shown in the figure above. (Fig.1)
• The case goes like this we have a pressure inlet at 100001 Pa, the outlet is at 100000 Pa.
• The domain has been initialised with 100001 Pa pressure with air as the specie having O2 and N2 concentrations reversed.
• The air with normal concentration is supposed to enter from the inlet and should leave the domain from the outlet as a result of pressure differential.
• This need to be solved for three mesh sizes 2e-4, 1.5e-4, 1e-4.

The case needs to be setup in the CONVERGE STUDIO 3.0 and needs to be solved by CONVERGE SCIENCE SOLVER package using command line interface of Cygwin and post processed to be viewed in PARAVIEW 5.4.

Solver details:

Steady state Pressure Based solver

Solver Scheme: PISO

Max PISO iterations: 20

Start time: 0 cyc

End Time: 15000 to 18000 cyc

dt start: 1e-9

dt max: 1

dt min: 1e-9

Equation solved: Momentum, energy, specie (mainly)

Initialisation:

Specie: N2 0.23 (For the User Defined Region)

O2 0.77

Pressure: 100001Pa

Temperature: 300k

Velocity: [0,0,0] m/s

Boundary Conditions:

Inlet: Type: Inflow

Pressure 100001Pa

Velocity Neumann

Temperature: 300k

Specie: air(i.e. O2 0.23 & N2 0.77)

Outlet: Type: outflow

Pressure 100000Pa

Velocity Neumann

Walls: No Slip wall

The results obtained are shown below:

Results as compared on graph:

Average Velocity on Outlet Boundary

Total Pressure

Cell Count

Inferences:

• The flow parameter gradually attains the steady state after stipulated number of cycles.
• For all the graphs above the curve for largest grid size shows the results which are completely different from the other two fine grid sizes.
• For all the parameter like pressure, velocity and mass flow rate the fine grid size go hand in hand.
• It can also be inferred that if suppose the case had some turbulence features, they could only be captured by the finest grid in detail and would also require large computation time.
• As we see that for average velocity the curve for grid size 1.5e-4 and 1e-4 cross each other at various points and so they show results that are more accurate.
• For the simple case a large grid size can be handy when one just needs perceive the flow characteristics in broader sense, but when it comes to performance analysis and improvement a small grid size is preferred, at the cost of computational effort and the physics involved need the stipulated attention.

The flow parameters for the various grid sizes in shown below:

@mesh size: dx=2e-4, dy=2e-4, dz= 2e-4

Average Velocity Contour

Final velocity (avg.): 3.44m/s (because inlet is considered negative and outlet as positive)

Total Pressure

Total Pressure: 100009(outlet)-100010(inlet)Pa (It maintains the difference of 1Pa through-out as given in the boundary condition)

Mass Flow Rate

Final Mass Flow Rate: 0.0427Kg/s (because inlet is considered negative and outlet as positive)

@mesh size: dx=1.5e-4, dy=1.5e-4, dz= 1.5e-4

Average Velocity Contour

Final velocity (avg.): 2.29m/s (because inlet is considered negative and outlet as positive)

Total Pressure

Total Pressure: 100004(outlet)-100005(inlet)Pa (It maintains the difference of 1Pa through-out as given in the boundary condition)

Mass Flow Rate

Final Mass Flow Rate: 0.028Kg/s (because inlet is considered negative and outlet as positive)

@mesh size: dx=1e-4, dy=1e-4, dz= 1e-4

Average Velocity Contour

Final velocity (avg.): 2.22m/s (because inlet is considered negative and outlet as positive)

Total Pressure

Total Pressure: 100004(outlet)-100005(inlet)Pa (It maintains the difference of 1Pa through-out as given in the boundary condition)

Mass Flow Rate

Final Mass Flow Rate: 0.027Kg/s(because inlet is considered negative and outlet as positive)

Mesh Contour

2e-4

1.5e-4

1e-4

Line Plot For X-velocity at 100^th cycle for various mesh

Conclusion:

This project helped me a lot in learning how to setup a case in CONVERGE CFD run a simulation onit using Cygwin command line interface and perceive the post processed results in PARAVIEW.

This project not only introduced but gave me an opportunity to know the capabilities of open source software like PARAVIEW and CYGWIN, and also an insight into how powerful and informative CONVERGE for a new comer like me.