CHT ANALYSIS IN AN EXHAUST MANIFOLD USING ANSYS FLUENT

OBJECTIVE:

The aim of this project is to perform a Conjugate Heat Transfer analysis in an exhaust manifold model using ANSYS Fluent. Also to study the effect of mesh refinement on the velocity, temperature and heat transfer coefficient(HTC) values of the flow by observing the appropriate plots and contours.

CHT Analysis:

A CHT analysis is used in describing the transfer of heat between a solid and a fluid medium due to thermal interaction at their interactive points. The conjugate heat transfer problem is governed by the set of equations consisting in conformity with physical pattern of two separate systems for body and fluid domains.

Application of CHT analysis:

1. Cooling of engines.

2. Water jacket efficiency calculation.

3. HVAC.

4. Race car aerodynamics

5. Turbo and super chargers

Need of CHT analysis:

Wherever a fluid and a solid domain comes into contact, there will be exchange of energy mostly in the form of heat. It should be predicted previously in order to efficiently design the system. For example, in an air-cooled engine, CHT analysis gives the area where the temperature and HTC is maximum. So its the area where cooling is much needed. So the design and mounting of engine is changed in that favour. Likewise, for all the applications mentioned above uses CHT analysis as a preventive and saviour plan.

SIMULATION PROCEDURE:

Fluid volume extraction:

The first step of the procedure is importing the exhaust model in spaceclaim. After removing the the unwanted entities, the fluid volume has to be integrated as the main aspect of the process is to study the heat transfer between the interacting fluid and solid domains. The solid and fluid volumes are as below.

Fig-1:Exhaust manifold model

Fig-2:Extracted fluid volume

Meshing:

After seperating the domains, the next step is to mesh the model. For this simulation, two mesh setups will be created. First one with the baseline parameters already been set by the software and the simulation is carried out. And the second one with a much refined setup with smaller mesh elements and with some special meshing methods incluuded.

Baseline setup parameters:

1. Mesh element size - 150 mm

2. Inflation - Nil

3. No. of nodes - 27484

4. No. of elements - 137926

5. Type of mesh element - Tetrahedral

The mesh created using the above mentioned baseline parameters is displayed below.

Fig-3:Baseline mesh setup

Refined setup parameters:

1. Mesh element size - 150 mm

2. Inflation in solid domain

1. No.of layers - 3
2. Growth rate - 1.2
3. Maximum thickness - 5 mm

3. Inflation in fluid domain

1. No.of layers - 3
2. Growth rate - 1.2
3. Maximum thickness - 5 mm

4. Body sizing method - Solid domain with 10 mm

5. No. of nodes - 106251

6. No. of elements - 322936

7. Type of mesh element - Tetrahedral

Fig-4:Refined mesh setup

The inflation layers are so important as it creates a body fitted mesh. Since this simulation majorly focuses on the interaction area, a body fitted mesh is much need to capture and study the characteristics of the flow field variables. The inflation layer is created as in the below figure.

Fig-5:Refined mesh setup with inflation layers

Physics setup and calculation:

Then these mesh models are solved in fluent seperately with the below mentioned parameters. There is no change in the physical setup of the problem. Both the models are solved with the same parameters.

1. Solver - Steady state, Pressure based solver

2. Physical model - Viscous model with energy equation

3. Viscous model - Standard k epsilon

4. Solid material - Aluminium

5. Fluid material - Air

6. Boundary conditions

1. Inlet velocity - 5 m/s
2. Inlet temperature - 700 K
3. Outlet gauge pressure - 0 pascal
4. Outer wall of solid volume - Convection with heat transfer coefficient of 20 W/m^2 K

7. Initialization - Hybrid

8. No.of iterations - 150

After setting these parameters, the calculation is done for 150 iterations. The residual plots of both simulations is as below.

Fig-6:Baseline setup residuals

Fig-7:Refined setup residuals

As it can be seen, except the energy equation, residuals of other equations are almost identical for both the cases. But the energy equation is converged much earlier in the refined setup than in the baseline setup.

RESULTS USING CFD-POST:

The post processing of the flow field variables is carried out in CFD-POST. As a first step. a cut plane is created in a position as in the figure below as it is the anticipated area with extreme values. It is due to the fact that it has the highest mass flow rate in the entire manifold.

Fig-8:Plane creation in post-processing

Temperature:

The temperature distribution as anticipated has the highest value at the curved area displayed below. This is due to high mass flow rate of hot air. Also fricition is high in that bent area contributing to the high temperature.

Fig-9:Temperature distribution in baseline setup

Fig-10:Temperature distribution in refined setup

Comparing the results of two setups, it is evident that the temperature distribution is finer and more accurate in the refined setup. The baseline setup is very coarse making it unrealistic. The implementation of inflation layers is the major advantage of the refined setup making it more accurate especially at the interacting areas. A single layer of interaction can be noted in refined setup which is accurate in nature also.

Velocity:

In the overall flow, the maximum velocity is obtained in the exhaust port area as shown in the below figure. It is also due to high mass flow rate in that particular area.

Fig-11:Velocity streamlines in refined setup

The velocity contour plot of baseline setup is very coarse as the temperature plot. It is actually worst in capturing the boundary conditions. The maximum velocity here is 37.3 m/s . But the fact that the velocity will be higher at the curvature region is depicted correctly in both the setups.

Fig-12:Velocity distribution in baseline setup

Fig-13:Velocity distribution in refined setup

The best thing about the velocity contour of refined setup is its ability to capture the boundary conditions accurately. Especially, the no slip boundary condition at the wall is captured only in the first layer of the flow which is achieved by the creation of body fitted mesh. It also looks way finer than the former one. Here the maximum velocity is 40.6 m/s.

Heat transfer coefficient:

The maximum value is achieved in the curvature area in both the setups which is right. But there is a huge difference between the maximum values of these setups. It is 100.6 W/m^2 K in baseline setup but 190.5 W/m^2 K in refined setup. It is nearly double the time than the former.

Fig-14:Wall heat transfer coefficient in baseline setup

In baseline setup, the HTC plot is very coarse as seen in the above figure making it inaccurate. Because heat transfer between two mediums will take place only at the first layer of interaction. This is exactly depicted using a well refined mesh.

Fig-15:Wall heat transfer coefficient in refined setup

Fig-16:Wall heat transfer coefficient in refined setup(closeup view)

The mesh pattern at the cut plane is as displayed. It clearly explains the accuracy of the refined mesh over the baseline setup.

Fig-17:Mesh in baseline setup

Fig-18:Mesh in refined setup

VERIFICATION OF HTC PREDICTION:

The heat transfer coefficient predicted by the simulation can be verified in two ways.

1. Using the local heat flux values and the wall adjacent temperature.

Since `h = ((dQ)/(dt) )/(A.DeltaT)` , it can be verified after finding the local wall temperature and heat flux using probes in post processing.

2. Using Nusselt's number.

We know that `h = (Nu.k)/L` , where k is the thermal conductivity of the fluid and L is the first layer thickness. To calculate Nusselt number, this formula can be used,

`Nu = 0.023(Re)^(4/5)(Pr)^0.4`, where Re is the reynolds number and Pr being the Prandtl number. The prediction value of HTC can also be verified by this method.

Factors affecting the accuracy of prediction:

1. Size of the mesh element.

2. Type of mesh used.

3. Number of mesh elements and its quality.

4. Presence of inflation layers.

5. Boundary layer conditions

CONCLUSION:

The conjugate heat transfer analysis highly depends on the mesh that has been generated. As specified above, the accuracy of prediction of HTC majorly depends on the mesh parameters. Not only for HTC but also for other flow field variables such as temperature and velocity. In the mesh generation process, implementation of a body fitted mesh is very much important if it is a CHT problem. The inflation layers should be there at the solid-fluid interaction areas to calculate the values acccurately.