Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Problem with convergence of solution in 2D U shape Thermal Stress
Posted 2016年3月30日 GMT-4 15:27 Structural Mechanics Version 5.2 2 Replies
Please login with a confirmed email address before reporting spam
I want to know how much deflection at some point in my structure in "Thermal Stress" module
So from the beginning, I chose 2D Structure Mechanics - Thermal stress, and stationary as physics.
I defined geometry (in micrometers), just four rectangular domains connected together, making something like "U" shape.
Then I defined material, polysilicon(default)
Next is considering boundary conditions and initial values.
For defining "Solid Mechanics", I created - "Fixed constraint" and I fixated 2 ends. of "U"
For "Heat transfer in Solids", I created - "Heat source" and chose "Overall heat transfer rate" 1mW in one domain. - I did this for assuming that a heating power of 1mW is uniformly dissipated along that domain.
Also, I did not forget to define all the thickness for 'Solid mechanics' and 'Heat transfer in solid', and choose "plane stress"
I created mesh, normal size for first.
Now, I clicked study.
However, I get this message
"Failed to find a solution
The relative residual (3.5) is greater than the relative tolerance.
Returned solution is not converged.
- feature: Stationary Solver 1 (sol1/s1)"
Am I missing something?
So from the beginning, I chose 2D Structure Mechanics - Thermal stress, and stationary as physics.
I defined geometry (in micrometers), just four rectangular domains connected together, making something like "U" shape.
Then I defined material, polysilicon(default)
Next is considering boundary conditions and initial values.
For defining "Solid Mechanics", I created - "Fixed constraint" and I fixated 2 ends. of "U"
For "Heat transfer in Solids", I created - "Heat source" and chose "Overall heat transfer rate" 1mW in one domain. - I did this for assuming that a heating power of 1mW is uniformly dissipated along that domain.
Also, I did not forget to define all the thickness for 'Solid mechanics' and 'Heat transfer in solid', and choose "plane stress"
I created mesh, normal size for first.
Now, I clicked study.
However, I get this message
"Failed to find a solution
The relative residual (3.5) is greater than the relative tolerance.
Returned solution is not converged.
- feature: Stationary Solver 1 (sol1/s1)"
Am I missing something?
2 Replies Last Post 2016年3月31日 GMT-4 08:21
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hello Hoon Jeong,
If the residual error is greater than the relative tolerance, means that your convergence will never meet.
What is your relative tolerance?
Perhaps try to increase your relative tolerance to 1, and then decrease to see where its limit is.
The higher your relative tolerance is, the lower your accuracy is, however, you get convergence really quickly. So use a relative tolerance of 1, and then decrease it slowly to see where your convergence limit is.
I hope that helps.
If the residual error is greater than the relative tolerance, means that your convergence will never meet.
What is your relative tolerance?
Perhaps try to increase your relative tolerance to 1, and then decrease to see where its limit is.
The higher your relative tolerance is, the lower your accuracy is, however, you get convergence really quickly. So use a relative tolerance of 1, and then decrease it slowly to see where your convergence limit is.
I hope that helps.
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hello Hoon,
What boundary conditions do you have on the heat transfer in solids equation?
By default, all boundaries are thermally insulated. That is not compatible with a stationary solution if a uniform non-zero distributed heat source is applied (The heat you put in via the distributed heat source must be able to come out somewhere on the boundaries - The divergence theorem applied to the heat flux), so you can't keep all default BCs.
Best,
Jeff
What boundary conditions do you have on the heat transfer in solids equation?
By default, all boundaries are thermally insulated. That is not compatible with a stationary solution if a uniform non-zero distributed heat source is applied (The heat you put in via the distributed heat source must be able to come out somewhere on the boundaries - The divergence theorem applied to the heat flux), so you can't keep all default BCs.
Best,
Jeff