Performing a CFD Analysis to determine the efficiency of the Mixing Tee for two different models using K-epsilon Turbulence Model in ANSYS Fluent.
PROJECT TITTLE: PERFORMING STEADY STATE SIMULATION ON A MIXING TEE GEOMETRY USING ANSYS FLUENT.
OBJECTIVE:
Setup steady state simulations to compare the mixing effectiveness.
- Case 1
- Short mixing tee with hot inlet velocity of 3m/s
- Momentum ratio of 2
- Case 2
- Long mixing tee with hot inlet velocity of 3m/s
- Momentum ratio of 2
- Case 3
- Short mixing with hot inlet velocity of 3 m/s
- Momentum ration of 4
- Momentum ratio = velocity at cold inlet / velocity at hot inlet
In a table, compare the following values from the 3 cases.
- Cell count
- Average outlet temperature
- Number of iterations for convergence
For each simulation, show the convergence plot.
Show a cut plane along the middle of the pipe and colour it by velocity and temperature. These plots should be made for all the three cases with uniform legends.
INTRODUCTION:
Mixing Tees utilizes a specifically engineered internal geometry to efficiently mix two fluid streams into one combined stream. Mixing Tees are a simple, cost effective method of providing chemical diffusion to a system.Mixing Tees are ideal for microbore or analytical gradient HPLC. These mixing tees are specifically designed for high pressure applications.
Let’s say, a person is sitting in an air-conditioned room. It is working at a temperature of 10-degree Celsius (optimum efficiency temperature). But that person is not willing to sit in such a cold atmosphere. Here comes the requirement of Tee mixer. It is called Mixing Tee because the two fluid paths are met at an angle of 90-degree. It is used as a flow control device. To get a soothing temperature around 22-26-degree Celsius, the chilled 10-degree air-conditioned air can be mixed with the hot air from the ambient using a Mixing Tee. This will result in a pleasant temperature at its output.
Similarly, when a person is taking a shower with the help of a hot geyser, the temperature might be too hot to bath. Here also one can implement a Mixing Tee with the normal tap or tank water mixed with the geyser water to get a relaxing shower.
NUMERICAL ANALYSIS ON MIXING TEE:
3D CAD MODEL:
The model is imported to Space Claim module. Now, the 3D CAD model has thickness. Had there been any heat flux applied to the Tee wall then we would have required this thickness into our solver i.e. ANSYS Fluent. The physics of this particular Mixing Tee problem does not require the thickness. We just need the wet surface area. Wet surface area means the area with which the fluid will be in contact with.
For this purpose, we need to extract the wet volume and prepare it for simulation, the rest of the geometry will be suppressed for the physics. At the top of Space Claim there is an option as ‘prepare’, under it ‘volume extraction’ is selected and all the three edges are marked and ‘right’ option is checked. This will lead to volume extraction made up by those three edges from the initial geometry. Once the volume is extracted, it is then opened in ANSYS Mesher for meshing.
The arrows are highlighting the edges that are selected for volume extraction. The figure given below is the extracted volume.
Meshing and Boundary definition:
The element size is 1.5mm. The domain is meshed with linear tetrahedrons. Here is how the mesh looks along with element quality.
After this, boundary names are assigned. This is done to easily put the boundary conditions in Fluent. After this Fluent is opened with ‘Contact Precision’ option checked in at the starting. Firstly, a mesh check is being conducted to see for any mesh quality abnormality. Once it’s done, the physics which governs the Mixing Tee simulation for our objective is checked in. Steady state approach because we are just interested in finding out what is the temperature at the Tee outlet, in between we do not intent to check. Pressure-based because the Mach No. for this problem is very less hence no negligible change in density. Velocity formulation is absolute. Since, we want the temperature distribution, therefore, energy equation is also solved. We also want to introduce the concept of viscosity into our problem. For this, additional equations will be solved, and viscosity will also influence the mixing region. K-epsilon(2 eqns.) viscosity model is implemented with ‘Realizable’ form. This is a modified form of k-epsilon.
For material selection, ‘Cell Zones’ is selected, and the volume is edited. There we will get the option of assigning the fluid or add and then assign. For this analysis, air is selected.
Now, assigning the boundary conditions, the outlet of the Tee is a pressure B.C with gauge pressure as 0 Pascal. The inlet chilled air has a temperature of 10-degree Celsius while the inlet hot air has a temperature of 25-degree Celsius. The velocities are given according to the cases in the objective as stated above.
At last, initialization is being done and the method used is Hybrid. Under definitions option, new surface reports are created for temperature and velocity at the outlet as Area-weighted average. A standard deviation report is also created for temperature to see the deviation of the temperature throughout the iterations at the outlet. I have made the simulation run for 150 iterations and recorded the plots and the data as mentioned in the objective of this project.
RESULTS AND DISCUSSION:
CASE I:
Cut plane showing the Temperature profile-
Cut plane showing the Velocity profile-
Residuals Plot:
Standard temperature deviation, Area Weighted Average Temperature and Velocity at Outlet:
CASE II:
Cut plane showing the Temperature profile-
Cut plane showing the Velocity profile-
Residuals Plot:
Standard temperature deviation, Area Weighted Average Temperature and Velocity at Outlet:
CASE III:
Cut plane showing the Temperature profile-
Cut plane showing the Velocity profile-
Residuals Plot:
Standard temperature deviation, Area Weighted Average Temperature and Velocity at Outlet:
CONCLUSION:
|
CASES |
CELL COUNT |
AVERAGE OUTLET TEMPERATURE(OC) |
NO. OF ITERATIONS FOR CONVERGENCE |
STANDARD DEVIATION OF TEMPERATURE |
|
Case I |
Nodes: 41721 Elements: 211462 |
20.28 |
200 |
1.99 |
|
Case II |
Nodes: 56727 Elements: 288361 |
20.37 |
350 |
1.43 |
|
Case III |
Nodes: 41721 Elements: 211462 |
17.8 |
200 |
0.908 |
From the above table, we can conclude that Case III setup provides the best mixing efficiency since its standard deviation is lowest. This can also be inferred from the temperature cutplane contour which clearly shows that the primary mixing at the junction of the two inlets is maximum from the other two cases, this provides better mixing due to turbulence which is also created by high velocity chilled air coming in. The convergence can be deduced from the residuals plot, the point at which all the residuals follow the same pattern onwards, that is the point where we can say that the solution has reached convergence and there is no need to compute further. You might wonder that the temperature and the velocity line plot became horizontal much earlier around 60 iterations, but the scale at which the residuals are being computed is very small and takes variations into account of the order of 1e-4. Therefore, if the velocity and the temperature line plot are plotted at this scale then we will see the deviation.