Exhaust Port Challenge
Objective: To perform Conjugate Heat Transfer Analysis on an Exhaust Manifold.
Conjugate Heat Transfer is a process in which Heat Flows through both solid and liquid which are in close contact with each other. The mode of heat transfer in solid is conduction and in fluid is convection. CHT analysis is important in case where there are two fluids exchanging heat with each other and they are seperated by a solid wall. The best example of such type of process is a Indirect contact Heat Exchanger. In case of a Heat Exchanger one fluid flows through inside of tube and other on the outside. First the heat is transferred from hot fluid to solid wall and then from the wall to the cold fluid. The term overall heat tranfer coefficient is used to calculate the heat transfer. In such cases CHT analysis is important to understand the effectiveness of heat exchanger.
In this Project the CHT analysis of an Exhaust manifold is done using ANSYS Fluent. The working fluid is air having an inlet temperature and velocity of 700K and 5 m/s respectively. The outer wall is set to convection heat transfer mode with Heat Transfer coefficient 20 W/m^2 K. Temperature and velocity variation is studied alond the manifold and variation of heat transfer coefficient is simulated.
1. Geometry
2. Mesh
Tetrahedral Meshes were used for both Fluid volume and solid volume. The mesh was refined by having a element size of 8mm for solid domain and 50 mm for Fluid domain. Inflation layer of total thickness 5mm and total 5 layers was added to accurately capture the boundary phenomonon.
Total Number of elements = 477432
Mesh
Mesh Metric
3. Results
The analysis was done using both k - epsilon turbulence model and k - omega SSt turbulence model and comparison between the two was done. In both the cases Steady state and Pressure based solver was used. The solver was run for 500 iterations and contours of temperature, velocity and heat transfer coefficient was plotted to understand their variation.
Case 1: K - Epsilon Turbulence Model
Residuals
Temperature variation at outlet
Temperature variation at outer wall
Temperature contour at oulet pipe
Streamlines
Velocity contour at oulet pipe
Wall Heat Transfer Coefficient
Wall Heat Transfer Coefficient at Bend
Wall Heat Transfer Coefficient at outlet pipe
By looking at the convergence plot we can see that there are total 7 equations being solved in the solver. These are continuity, Momentum (x,y and z), Energy, K and epsilon. The K term here is turbulent kinetic energy whereas the epsilon is turbulent dissipation Rate. The maximum value of heat transfer coefficient is 212.4 W/m^2 K.
Case 2: K - Omega SST Turbulence Model
Residuals
Temperature variation at outlet
Temperature variation at outer wall
Temperature contour at oulet pipe
Streamlines
Velocity contour at oulet pipe
Wall Heat Transfer Coefficient
Wall Heat Transfer Coefficient at bend
Wall Heat Transfer Coefficient at outlet pipe
In the second case also a total of 7 equations are solved. The only difference here is that instead of epsilon we have omega as turbulent dissipation. The maximum value of heat transfer coefficient in this case is 237.3 W/m^2 K.
Comparison between K-epsilon and K-omega model
|
Parameter |
K – ε model |
K – ω model |
|
No. of iteration to converge |
200 |
350 |
|
Outlet Temperature |
676.8 K |
677.5 K |
|
Maximum Temperature |
700 K |
700 K |
|
Minimum Temperature |
505 K |
484 K |
|
Maximum Velocity |
40 m/s |
40.62 m/s |
|
Maximum HTC |
212.4 W/m2 K |
237.3 W/m2 K |
After comparing the two turbulrnce model we can see that the k-omega model requires more number of iterations to converge. It is known that the k-omega SST model predicts the boundary phenomenon such as wall shear stress, drag, and in this case wall heat transfer more accurately. This can be seen by comparing the values of HTC for both the cases. The HTC is higher in k-omega model and the Minimum temperature is low.
Now as the two pipes on either side of inlet section merge we get an increase in velocity and thus by Dittus Boelter equation there should be an increase in the HTC. This is well predicted by the k-omega model as can be seen by comparing the two contours.
k-epsilon k-omega
So for this case we can say that the k - omega model is better in analysis of Heat transfer coefficient.
4. Verification
There are correlations available in the literature to calculate the Nusselt Number. One such correlation is the Dittus Boelter equation which is for internal flows having Reynolds number greater than 10000.
`Nu = 0.023*Re^(4/5)*Pr^n`
where Re = reynolds number
Pr = Prandtl number
n = 0.4 for heating
= 0.3 for cooling
So it can be seen that the Nusselts number is directly propertional to Reynolds number and we can say that HTC is directly proportional to velocity, since
`Re = (rho*V*D)/mu` and `Nu = (h.D)/k`
In our simulation the velocity is highest in outlet section and specifically at the bend. We can see that the HTC is also highest at the same place. So we can say that the variation in heat transfer coefficient are accurately predicted.