Iterations stop at save point with transient thermal analysis

[simon]

Hi,

I am trying to do a quick and dirty transient thermal analysis for a simulated weld joint. I can get convergent with a steady state just fine but when I try and run a transient analysis (our weld time is only around 30sec so not even close to steady state) the first couple of iterations go fine but then at the save iteration point the CPU% drops to 0 and the analysis seems to stall. When I run the sim with automatic convergence detection on at the same point it typically stalls it will say that convergence and been reached. This doesn't make sense to me as the "converged" solution from the transient analysis isn't even close to the steady state and happens at at more or less 1sec after the start of the weld. I am guessing that something is wrong with my set up or my solve settings. I have a .1sec time interval 5 iterations per step and a save point every second with an end time of 30. I would like to think i have all my boundary conditions set up right (I am currently in my last year of mechanical engineering) but I could use some input if you can.

I am currently at school and not on my CFD machine so I will post my files later this evening. Anything to get me started would be appreciated though.

Thanks,

Simon

[simon · Accepted solution]

Problem Solved: I moved the file location from my solid state drive (107GB Free) to my spare magnetic drive (1.71TB Fee) and now everything is working great with the results very close to what is expected. I can't explain this as the file after the render is only 12GB but perhaps someone else can help me understand. FYI the model is 35k solid nodes and 1.54m solid elements.

[André.Silva]

Maybe when it stops, it deletes all the time steps calculated.

When the simulation stops can you see all the time steps? If not then they were deleted...

Also you can reduce the overall simulation memory size by increasing the "Save intervals" from 1 to 5 or 1 to 2. This means that when you have 100 time steps but you save all the time steps and you have 50 gb then if you increase the "Save intervals" from 1 to 2 you should have half the memory (25gb) at 100 time steps. 

But of course, with this you won't be able to see all the time steps. This will depend on the objective of your simulation and what you want to analyse.