Exhaust Port Challenge

OBJECTIVES

1. Discuss why Conjugate heat transfer is used!.

2. To simulate the Exhaust Port in fluent with maintaining the y+ value according to the turbulence model.

3. Calculate the wall heat transfer coefficient on the internal solid surface & show the velocity & temperature contours in appropriate areas.

INTRODUCTION

Conjugate heat transfer defined as the heat transfer between two systems or domains (solid and liquid) by the exchange of thermal energy. Temperature defined the amount of thermal energy available and the movement of thermal energy determined by the heat flow. Heat transfer is solids occurs by conduction while in liquids it occurs by convection and at solid-liquid interface heat transfer occurs by convection. CHT refers to the heat transfer between solid and fluid boundaries by exchanging thermal energy between them.

CHT provides temperature distribution and hot spot region at the solid-liquid interface and also can predict heat transfer accurately, for example, conduction through solids, free and forced convection in the gases/fluids, and thermal radiation. It also provides the velocity and pressure distribution of the fluid inside the solid. We can use CHT in design optimization for improvement in heat transfer rate and cooling capacity.

Conjugate Heat Transfer used in turbochargers, heat-exchangers, engine equipment, Heating Ventilation & Air condition system, Thermal Management of electronic equipment, exhaust ports, etc.

Exhaust Port

In exhaust port, fluid comes from the engine through inlets of the port that have a high temperature and when it passes through exhaust port so it will transfer the heat from the fluid to solid wall. Then it will increase the temperature of the solid surfaces and transfer heat to the environment. The process of heat transfer at a liquid-solid interface called the Conjugate Teat Transfer.

Methodology

1) Modeling

2) Meshing

3) Pre-processing

4) Setting up physics and solving

5) Post-processing

Software used:

a) ANSYS SpaceClaim: Geometry Modeling and repair

b) ANSYS Mesher: Generate a specified Mesh

c) ANSYS FLUENT: Physics set up and solving

d) ANSYS CFD Post: Post-processing 

1) Modeling: 

 Create a 3d model in ANSYS Space claim or create it in any other cad software and import it to ANSYS Spaceclaim.

Here I just import it.

Repair all the extra edges and extract the volume.

Now we have a 2 region a) solid b) fluid. Just unable the shared topology so that information passed from solid to liquid or vise-versa.

Also unable the share option present in workbench. This will share the data of the common surface.

This purple line indicates that geometry is ready for simulation.

CASE 1: BASELINE MESH (Element size 150)

2) Meshing:

Define the names of the boundaries

 Start with the baseline mesh to check the solver first. Generate mesh with the element size 150 mm.

Also, take care of the solid-fluid interface and check the mesh so that the transfer of data between them.

3) Setting up physics and solving

for this step lets open up the Fluent and first check the Mesh. If Mesh is accurate then define the solver(Pressure Based solver).

Here we take Steady-state condition with K-epsilon turbulence model with standard wall function. Also, enable the energy equation.

Inlet velocity = 5 m/s

Inlet Temperature= 700 k 

Heat Transfer coefficient of the outer wall convection = 20 W/m^2-K

After providing these Boundary Conditions, start the simulation with hybrid Initialization then run for 350 iterations.

This graph shows the residual change.

Temperature contour of the exhaust port.

Velocity streamlines and velocity vector inside the exhaust port.

Temperature variation at the outer wall.

Wall heat transfer coefficient near the outlet.

Pressure Variation near the outlet.

Temperature variation near the outlet.

Velocity Distribution near the outlet. We can see that it is very high near the turning point.

In an exhaust post analysis, we assume that fluid flow through the manifold is too hot and transfer the heat from inside to the external surface. And we take a particular heat transfer coefficient of air near the exhaust surface that depends on the Re of external air. so this set up Provide a heat transfer coefficient when heat transfer from the fluid to solid and with the help of this we find out the heat transfer rate.

Since we run the set-up for a baseline mesh we get the expected results as shown above but as we study on the velocity contour and wall heat transfer coefficient, we get some diffused solution because of the unstructured mesh near the wall. In computational fluid dynamics, we are very sensitive to selecting the correct meshing which mainly depends on the element size. Normally we did a mesh sensitivity study and after getting the appropriate residual graph for convergence we take it for the problem.

CASE 2: Refined MESH (Element size of solid 16 mm & add some inflation layers)

 In refined mesh, we include the inflation layers to find out the result close to exactly near the walls. The height or thickness of the inflation layers depends on the different turbulence models. In this case, if we go for high Reynolds number then we choose the k-elipson turbulence model with standard wall function which y+ range will be 30-300 and while we use k-omega SST, the y+ range should be 1-5. In some cases, it goes below 1 to calculate more information near the walls.

For the refinement of the case, we provide the inflation layers. We start with the initial guess (Total thickness of 5 mm). We can increase or decrease the total thickness it will totally depend on the y+ range. If the y+ range exceeded or decrease from the desired limit we should change our initial guess.

Also, we provide the edge sizing of the solid (i.e 16 mm) for the refine mesh but should take care of the maximum number of cells because in the academic version I will not go above .5 million cells. Also, turn on the capture curvature to create proper mesh on the curves.

Nodes: 162750

Elements: 462644

We should also take care of the region between the solid and liquid. So it is necessary to turn on the shared topology so that easily transfer of information between them. Otherwise, we should not get the correct result.

3) Setting up physics and solving

for this step lets open up the Fluent and first check the Mesh. If Mesh is accurate then define the solver(Pressure Based solver).

Here we take Steady-state condition with K-epsilon turbulence model with standard wall function. Also, enable the energy equation.

Inlet velocity = 5 m/s

Inlet Temperature= 700 k 

Heat Transfer coefficient of the outer wall convection = 20 W/m^2-K

After providing these Boundary Conditions, start the simulation with hybrid Initialization then run for 350 iterations.

This graph shows the residual change.

Temperature contour of the exhaust port.

Velocity streamlines and velocity vector inside the exhaust port.

Temperature variation at the outer wall.

Wall heat transfer coefficient near the outlet.

Pressure Variation near the outlet.

Temperature variation near the outlet.

Results

We perform the simulation for both the cases and get the Velocity, temperature, wall heat transfer contours. We can see easily that the refine mesh contours get more exact results as compare to the baseline mesh diffused results. We start with a baseline mesh to check the settings and run the solver. Once we get the result then apply the finer mesh and get the solution. In this case, we get the converged solution and get the information of velocity and temperature close to exact.

Observation Table:

From the above observation, we can say that refine mesh produces more accurate results near the wall due to proper interpolation.

Validation of Results get from Simulation:

We can validate the result get from the simulation in two ways. Either we take the data from someone's research paper who already done this type of work or we set up an experiment to perform this analysis and validate the result.

The following factors affect the accuracy of the Heat transfer prediction

1) Mesh: Choose correct mesh (Including inflation layers and edge sizing) which will give the accurate prediction of wall heat transfer coefficient near the walls.

2) Turbulence Model: Choose the Turbulence model wisely according to Reynold's number and y+ value.

                               For Low Re choose k-omega SST Turbulence Model (y+ 1 to 5)

                               For high Re choose k-epsilon with standard wall function Turbulence model (y+ 30 to 300)

We can also get a theoretical heat transfer coefficient with the help of nusselts number.

`Nu = 0.023 * Re^(4/5) * Pr^0.4`

`Nu = frac{hl}{k}`

Inference

We used the conjugate heat transfer analysis for find out the heat transfer between solid and liquid surfaces. It is widely used in various applications such as automobils, heat exchangers, turbochargers. We used CHT analysis here on the exhaust port to find out the wall heat transfer coefficient and velocity near the region of the wall. And we start with the baseline mesh to check the setup and solver first and after that refine the mesh to get the result close to exact near the walls. Mesh quality and right turbulence model play an important role to give the prediction of the close to exact result near the wall.