Page History

...

 <iframe width="640" height="360" src="https://www.youtube.com/embed/pyvOziL5mso" frameborder="0" allowfullscreen></iframe>

From Prior Simcafe Version:

Verification and validation can be thought of as a formal process for checking results. Each of these terms has a specific meaning which we won't get into here. The only verification check discussed below is investigation of the effect of mesh size on the results.

Refine Mesh

Let's check what happens to the heat flux variation along the bottom surface when we refine the mesh i.e. use more elements in the mesh.

 <iframe width="600" height="338" src="//www.youtube.com/embed/F1_I9oMJ1XE" frameborder="0" allowfullscreen></iframe> 

Summary of the above video:

1. Duplicate Model branch in project page
2. Bring up Mechanical window by double-clicking on Model and double edge sizings. Since we double the number of divisions in x and y directions in the refined mesh, the number of elements is increased by a factor of 4 (2x2) as compared to the original mesh.

To resolve the steep temperature gradient at right, lower corner, we bias the edge sizings using the following process:

 <iframe width="600" height="338" src="//www.youtube.com/embed/WGeYMRORW4A" frameborder="0" allowfullscreen></iframe> 

Summary of the above video:

1. Check direction of edges by clicking on arrow in top toolbar.
2. For first edge sizing, select the bias type and apply bias factor of 10. Rescope to just the bottom boundary. Note that the bias factor is the ratio of the largest division to the smallest division.
3. Repeat for second edge sizing, rescoping to left boundary.
4. Duplicate these edge sizings, rescope to upper/right boundaries and reverse bias type.
5. Set Behavior = Hard for all four edge sizings so that ANSYS does not mess with the edge sizing inputs provided. (You don't need to repeat this if you already did this in the meshing step.)

To update prior results on the new mesh, click Solve. Highlight the column in the tabular data containing the updated heat flux values along the bottom surface and copy into the Excel file in a new column.

We repeated the above process for a third mesh obtained by doubling the edge sizings on the second mesh while keeping the bias factors the same. The following figure shows that the heat flux variation along the bottom surface is almost the same on the two finer meshes (20x40 and 40x80). This demonstrates that the results on the two finest meshes are almost mesh independent and so are likely to be very close to the true solution.
Image RemovedThe above figure was generated in MATLAB from the Excel file using this MATLAB script. The Excel file called boundary_flux_tutorial.xlsx had four columns as shown below.
Image RemovedGo to Exercises

Go to all ANSYS Learning Modules