Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
Too large thermal expansion with a small temperature variation
Posted Feb 18, 2015, 5:43 a.m. EST Heat Transfer & Phase Change, Studies & Solvers, Structural Mechanics Version 4.2, Version 4.4 1 Reply
Please login with a confirmed email address before reporting spam
As the title says. My temperature varies only on the scale of a few Kelvins, but still I get a engineering strain of more than 1. I am trying to model an IC engine cylinder made out of cast iron for a school project. The temperature inside varies as a periodic function of time, resembling the movement and ignition of piston and fuel. I have made some considerations for the boundary conditions and symmetry which are included in the pdf file in section 2. I have also included the mph file itself and a diagram of the problem in 3D.
I have observed the following.
1. The strain is far too large to make any sense.
2. The solver makes timestepes of 1sec even so the heat influx varies in the millisecond scale. I have tried forcing the solver to make one ms timesteps, but the problem persists.
3. Under certain conditions the cylinder is displaced across a fixed contrain.
4. I have tried reducing the aspect ratio and decreasing the mesh size. No effect.
5. I have tried solving the same problem on a computer with v.4.4. Same problem.
I would be thankful for any ideas/suggestions as I am currently drained of any ideas for solving this problem.
I have observed the following.
1. The strain is far too large to make any sense.
2. The solver makes timestepes of 1sec even so the heat influx varies in the millisecond scale. I have tried forcing the solver to make one ms timesteps, but the problem persists.
3. Under certain conditions the cylinder is displaced across a fixed contrain.
4. I have tried reducing the aspect ratio and decreasing the mesh size. No effect.
5. I have tried solving the same problem on a computer with v.4.4. Same problem.
I would be thankful for any ideas/suggestions as I am currently drained of any ideas for solving this problem.
Attachments:
1 Reply Last Post Feb 18, 2015, 2:36 p.m. EST
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Hi
I have V5 so your model is loaded with warnings, therefore I'm not sure that some of the issues I see are not coming from some translation issues, but:
1) you impose a very short and steep temperature variations these are not easy to catch, so you must guide better your solver with time steps adapted to your imposed profile.
2) you ask for 100 steps from 0 to 10 second but the temperature profile seem to last at most 10 msec, here you have a scale issue
3) the default time stepping settings are for typical diffusion problems, COMSOL cannot know that you are demanding short highly varying BC values, so you must change the default solver settings (deep down in the solver sub-nodes) to at least "intermediate" time stepping, and not the default "free"; or even "strict", and then make a far longer range of values really scanning your true BC temperature profile, including stepping closely around your steep slope transitions to help the solver to follow the transitions you impose
Then finally, I would use a mesh somewhat finer, with at least 3 elements across the thickness of your wall
--
Good luck
Ivar
I have V5 so your model is loaded with warnings, therefore I'm not sure that some of the issues I see are not coming from some translation issues, but:
1) you impose a very short and steep temperature variations these are not easy to catch, so you must guide better your solver with time steps adapted to your imposed profile.
2) you ask for 100 steps from 0 to 10 second but the temperature profile seem to last at most 10 msec, here you have a scale issue
3) the default time stepping settings are for typical diffusion problems, COMSOL cannot know that you are demanding short highly varying BC values, so you must change the default solver settings (deep down in the solver sub-nodes) to at least "intermediate" time stepping, and not the default "free"; or even "strict", and then make a far longer range of values really scanning your true BC temperature profile, including stepping closely around your steep slope transitions to help the solver to follow the transitions you impose
Then finally, I would use a mesh somewhat finer, with at least 3 elements across the thickness of your wall
--
Good luck
Ivar