exhaust port challenge

Conjugate heat transfer analysis:

As the name suggest it is the coupling of two simulations.

This analysis allows computation of the heat transfer between solid and fluid domains by exchanging thermal energy at the interface between them. Coupled boundary conditions are available for wall zones that separate two cell zones. Either the solid zone or the fluid zone, or both, may contain heat sources.

In solid region we are simulating heat conduction when it comes in contact with the fluid and in fluid region we are simulating heat convection (forced or natural ), where thermal energy are transfered due to movement of fluid. And at the interface between these two domain we have thermal coupling,therefore it\'s name conjugate heat transfer because we are considering the mass and thermal energy transfer at the same time
both in solid and fluid.

• Advanatages:

• Accuracy and speed.
• Sutitable for complex scenarios.
• No limits on no. of region to simulate.
• All combination of region types (solid-solid,solid-fluid,fluid-fluid).

• Applications:

• powerplant industries
• cooling of electronics devices
• refineries
• automobiles
• heat exchanger
• HVAC (heating,ventilating and air conditioner)
• combustion of burner
• gas turbine design
• engine cooling
• home appliances
• The below example show the temperature profile and streamline around a heat sink cooled by forced convection.

source : www.COMSOL.com

Objective: To perform a CHT on exhaust manifold model and to examine the velocity, temperature and heat transfer coefficient at the outer wall.

Imported the exhaust manifold model in spaceclaim and extracted the fluid volume. For proper transfer of heat between the fluid and solid volume, it need to have a conformal mesh at the interface of this volumes. Using share topology we can obtain a proper conformal mesh between the volumes.

Case 1: Basic mesh

The below image show the mesh formed using basic mesh setting.

All the element quality are above 10% and no. Of element is less than half million. The fig. below shows the metric chart and no of element.

For boundary condition we are creating named selection in the volumes.

4 inlet, 1 outlet and outer wall convection.

In the solution cell, before running the steady state simulation a mesh quality check was done to check the error. This is a pressure based problem as the mach no is below 0.3. To model heat transfer energy equation is activated. For viscous model, k epsilon turbulence model is used.

For this problem the material for solid is aluminium and for fluid is air.

Boundary condition:

Inlet= 5 m/s Temperature= 700K

Heat transfer coefficient (h) = 20 W/m^2 K (obtained form experiment)

Temperature contours along with animations are created to see the heat transfer.

This simulation was setup for 150 itirations.

From the residual graphs it can be seen that the simulations converged before 150 iteration . This convergence can also be examine by observing the animations and the temperature contours where the temperature doesn\'t change after a point of time.

For further analysis , we go to CFD post. Here, we obtain the temperature variations at the outer wall and find the maximum temperature . And, also the maximum velocity and heat transfer coefficient.

Temperature contours

Velocity contour

From the above contours by observing we can understand that the temperature appears to be high at bent near the outlet due to high velocity of fluid.

Heat transfer coefficient at the wall

In order to conserve mass,the velocity at outlet will be more than velocity at inlet. Because of this high velocity of fluid near the bent ,there appears to be high heat transfer coefficient near the wall which is shown in the image above.

Case 2: Refined mesh

Mesh settings:

The solid volume was meshed using a body sizing of 16mm along with capture curvature option to mesh the curved surfaces properly. To capture heat transfer coefficient at the wall and fluid interface inflation layer was used. A total of 5 layers was taken and the total thickness was 5mm.

Mesh with Inflation layers

The No. of element was below half million.

Since the mesh is fine so the simulation was set for 150+ iterations.

From the residual graph it can be seen that the all residual seems to converge after 140 iterations. This convergence can also be verified by observing the animations and temperature contours.

Temperature contours

Velocity streamline

Velocity contours at the plane

Heat transfer coefficient at the wall

Result and conclusion:

Case 1 Case2

• Velocity near the bent (m/s^2): 37.87 40.99
• Wall temperature at the bent(K): 582.2 602
• Heat transfer coefficient near: 110.9 217.1

From the above cases it can be seen that there is a difference in all the three parameters after refining the mesh.

The change in mesh can also be seen in the result of velocity contour at plane(basic mesh) where the velocity colour gradient at the bent look more diffuse as compared to refine mesh.

Case 2 gave us an idea that refining the mesh and adding inflation helps in predicting correct heat transfer coefficient simulation, which can be understood by looking at the above table.

Temperature contours for basic mesh

Temperature contours for refined mesh

VERIFICATION :

The result can be verified using Nusselt number, a dimensionless number which is the ratio of convective to conductive heat transfer.

By examining the velocity contours at the plane we can understand that there is high velocity at the bent which implies high reynold numer. From the above relation, we can say that as Re increases Nu increases as well, which inturn increases the heat transfer coefficient.

Heat transfer coefficient accuracy depends on:

• If we have conformal mesh between the volumes.
• Including the inflation layer for capturing heat transfer coefficient.
• No. of element used.
• Viscous model selection.
• Heat transfer coefficient for the outer wall (which is obtained from real experiment).
• conduction effects in solids.