Channel flow simulation using CONVERGE CFD

Aim:-

• Set the channel flow case setup in the converge.
• Run the base setup with three different base mesh sizes i.e. run 3 simulations and compare the results between all the 3 simulations.
• Plot velocity and pressure contours for all the 3 base mesh.
• Show mesh (i.e surface with edges) for the 3 base mesh size.
• Plot for velocity, pressure, mass flow rate, and total cell count for all 3 base mesh sizes.
• Explain the plots and write a detailed report on the project.
• Create animation in Paraview. Explain the development of flow using Paraview

Objective:-

• Create the channel flow geometry in the converge using the create option.
• Define the boundary to each surface of the geometry.
• Set normal direction on the inward side of the surfaces.
• In case of setup define the material, simulation parameters, fluid region, boundary condition, no of cycles, grid definition.
• Extract the setup files in the simulation folder.
• Run all the setup file Cygwin.
• Post-process all the results and open in Paraview.
• Show mesh for all 3 mesh bases.
• Plot velocity and pressure contour for all 3 base mesh sizes.
• Plot for velocity, pressure, mass flow rate, and total cell count for all 3 base mesh sizes.
• Create animation in the Paraview to show the development of flow.
• Explain the plots and write a detailed report on the project.

Geometry Creation:-

To start the project, as we require a geometry model to simulate, channel geometry was created using the shape -> Box option from the Create panel present in converge.

Setting up the normal:-

By default, the normals are facing the outside direction of the geometry so the Transform panel was used. Then by selecting any plane and hitting apply one can change the normal direction.

Boundary Definition:-

Boundary names were provided to each surface to assign proper boundary conditions to each face of the channel flow model.

Simulation Setup:-

The case setup was created after setting up the geometry and boundary definition.

Application Type:-

To perform normal steady-state simulation in setup application type we have to select a time-based option.

Material:-

In the material section 'Air' as fluid was considered.

Species:-

Species were defined as O2 and N2 as air is a mixture of Oxygen and Nitrogen.

Simulation Parameters:-

Run parameters:-

In the run parameter, the solver setup as the steady-state solver is defined. And in misc. Section the steady-state monitor was turned on to monitor the steady-state velocity for convergence of the cycle.

Steady-state monitor:-

Velocity monitor is set up with proper tolerance to keep track of velocity for convergence of results.

Simulation time parameters:-

General setting:-

Start time: 0

End time: 15000

Initial time-step: 1e-09

Minimum time-step: 1e-09

Maximum time-step: 1

Maximum convectional CFL limit: 1

Solver parameters [steady-state]

Steady-state solver control setting:-

Initial PISO convergence criterion multiplier (relaxed): 0.5

Boundary condition:-

Boundary:-

Inlet

For inlet boundary type as inflow selected with 100001Pa pressure and 300k temperature. And air species condition was specified by specifying O2 and N2 percentage as air is a mixture of 0.23% Oxygen and 0.77% nitrogen.

Outlet

At the outlet, boundary type OUTFLOW was selected, with outlet pressure 100000Pa and temperature as 300K. and air species condition was specified the same as an inlet.

Top and Bottom Wall

Top and Bottom faces were defined as walls with No-slip Condition and 300k temperature.

Side Walls:-

Left side ---> TWO-D

Right side ---> TWO-D

Initial conditions and Events:-

Regions an initialization

The initial fluid region was specified as air with 100000Pa pressure and 300K temperature and named as ‘flow volume’.

Grid control:-

The case setup ran for 3 different grid sizes. As the 2d mesh grid size along Z direction doesn’t matter much in simulation so it was kept as 0.01. It helps us to minimize the number of cell counts without affecting the results.

3 different mesh size are

For 1^st case

dx = 2e-4m, dy = 2e-4m, dz = 0.01m

For 2^nd case

dx = 1.5e-4m, dy = 1.5e-4m, dz = 0.01m

For 3^rd case

dx = 1e-4m, dy = 1e-4m, dz = 0.01m

Output/Post-processing:-

The time interval for writing 3D output data files kept for 100 cycles i.e. after 100 cycles it will create one 3D output file.

The time interval for writing restarting output was also kept for 100 cycles which will help us restart the simulation in the last count of 100 cycles in case simulation stopped unexpectedly. It helps to save the simulation time of the user.

Export Input files:-

Using file -> Export -> Export Input files all the input files exported to the simulation folder.

Simulation Run:-

The simulation was run with the help of Cygwin. The directory contains all the export files that were located and using the ‘converge.exe’ setup the simulation was run.

Results:-

Velocity contour:-

For 1^st case

For 2^nd case

For 3^rd case

From the observation from each velocity contour, we can find, each simulation providing almost the same velocity plot the only change in result contours is their resolution. the fine grid gives a smooth flow contour than the coarse grid but with a fine grid, computation time increases.

Pressure contour:-

For 1^st case

For 2^nd case

For 3^rd case

As velocity contour, the pressure contour of all three cases shows the same results except the difference in resolution of the picture.

Velocity Plot:-

For 1^st case

For 2^nd case

For 3^rd case

From the above velocity plot, we can say that the higher grid resolution improves the smoothness and accuracy of the results.

Velocity plot for the number of cycles:-

For 1^st case

For 2^nd case

For 3^rd case

The convergence factor depends upon the velocity motor tolerance. Therefore, all three simulations stopping after overcoming the provided tolerance to the average velocity value.

Pressure Plot for number of cycles:-

For 1^st case

For 2^nd case

For 3^rd case

The above figures depict the total pressure, over the period of cycles, increases and remain above 100000Pa.

Mass flow rate at Outlet plot for the number of cycles:-

For 1^st case

For 2^nd case

For 3^rd case

Mass at the outlet is always considered negative. all three simulations are showing very slight changes in mass flow rate.

Cell count plot for the number of cycles:-

For 1^st case

For 2^nd case

For 3^rd case

As there is no moving geometry, we used static mesh for the simulation where the cell count remains constant throughout the simulation cycles for all three cases.

Observation Table:-

With all the grid sizes we are getting the same results, which means the setup is correct and simulation results are dependable. Thus, results pass the grid independence test.

Animation of Development of flow:-

https://youtu.be/n7rT2zHNADQ

Animation exhibits the development of flow velocity as per the increment of time steps. as we can see shear stress acting on the fluid near the upper and lower boundary of the channel. Therefore, one can notice high velocity at the center part of the channel along the X-Direction and low velocity near the wall surface.

We can observe this difference more effective way with the help of the following graph.

Development of Velocity flow over a distance of X-direction:-

The above figure portrays the flow velocity development along with the distance of the X-direction at the final time step of the converged solution. here we can observe at the initial stage velocity at starting of the channel is the same form center to the wall. As flow move forwards the flow resistance gets applied by a wall over the fluid surface of the fluid. and then because of the viscosity effect, the shear stress gest applied to another layer of the fluid. hence, the flow velocity forms a curve as it moves along the X-direction, maintaining maximum speed at the center of the channel.

Conclusion:-

• Simulation with low grid resolution simulates the results with minimum time but gives results with little less accuracy and smoothness. Therefore to get an overview of the simulation it is the best choice to run the first simulation with coarse mesh size.
• Simulation with a fine grid preferably uses to perform the final simulation by using the same proven setup.
• simulating the results with different grid sizes is very important to check the mesh independence test.
• Flow velocity remains high at the center and low near-wall surface because of wall surface shear stress.
• If we plot velocity over the cross-sectional area of the channel we will get a parabolic profile curve on the graph because of the viscus effect of the fluid.