Caedium sometimes hangs on "Run Volume Solver."
Hello. I have a simulation which I saved after it was solved. When I open it back up and try to change certain physics parameters and re-run it to see the effects of the new parameters, the log shows that something is happening with the mesh, but then it says "End.," and the status simply says "Run volume solver." However, there is no way Iof which I am aware that I can then get the solver to run [again]. My thinking is that this is because the solver already ran the simulation before I saved it. However, I do not know how to see how the physics changes affect the simulation.
How can I do this?
Is there a way to save a simulation that will allow me to go back to it and make many changes to the physics and then re-run the simulation? I am trying to optimize a design and need to be able to make changes so I can. Thank you for your time.
Feedback
Questions? Ideas? Problems?
Create Residual Monitor
I will presume that you have successfully completed a tutorial, such as "Incompressible Flow Through a Pipe into a Box". If not I suggest you do that before any further simulations to get an idea of how Caedium works.
I'm guessing that you have not created a Residual Monitor - all the tutorials create it because it's important. Without it you can't tell what the solvers are doing. Assuming you are trying to run a steady-state simulation the traces on the monitor should reduce with increasing iterations. Ideally you want them to all be below 0.001 for the results to be converged. Also you should keep an eye on your results, e.g, contours of velocity magnitude. If you see unrealistic values (e.g., > 100 m/s for a inlet/free-stream of 10 m/s) then there is likely a problem with the mesh or the physics setup.
Before making changes to a simulation you should reset it by clicking Home->Start. Otherwise if the mesh is regenerated the simulation will immediately run until it reaches the same point it was at prior to the change. I think this is what you are seeing. With a Residual Monitor you'd see the solvers iterating. For more on the simulation controls see "Run, Stop, and Lock Controls".
Thanks for your reply. I have
Thanks for your reply.
I have done a number of the tutorials, and I have done the one with the incompressible flow from a pipe into a box quite a few times. I have also been creating a residuals monitor and using it to check for convergence.
What I am trying to do is to follow a simulation to convergence (or close to it) and then be able to go back to the physics of the simulation and change something, such as an outlet's velocity, and then have the simulation run again to reflect the effects of the change. Is there a correct way to do this?
However, when I try this via using the simulation's start button, it seems that the solver then just does not run, or at least I see no evidence of iterations happening. This is corroborated by my computer's activity monitor, which shows that Caedium is only running at a few percent of the CPU's load and with no evidence of simpleFoam running at all. When iterating, Caedium might only be running at a few percent, but I have 3 cores running simpleFoam at ~ 100% each, and the computer's fan starts within maybe 10-20 seconds and then runs constantly until the simulation is finished. Also, when I try to do this, Caedium seems to get stuck and I am unable to quit Caedium normally. If I then pull up my Mac's "Force Quit" dialog box, it says Caedium is not responding. I know that if Caedium is busy, it will say that in my activity monitor window for a few seconds, but then it is clear that Caedium is just working really hard, using ~100% of the CPU.
If I try to do this by using the "previous" button instead of the start button, it seems like I can make a physics parameter change, but then after the simulation runs to the end, the result looks the same. I understand that it might be because the change doesn't significantly affect the results, but when I go back to see if the simulation incorporated the physics parameter I changed by pulling up the properties for the changed item, I see that it hasn't because the previous value is back again; it has merely run the previous simulation from the start to the designated number of iterations again.
All this said, I am very much enjoying Caedium overall, and thanks in advance for your help!
Start is reset
The Home->Start button means set the simulation to the beginning (as in the start of a race). Then you use Home->Run to run the simulation again.
I suggest you click Home->Start (as in reset) prior to making changes. Once finished with your changes click Home->Run to run the simulation. As described in the tip "Run, Stop, and Lock Controls" you should only use Home->Previous to restore a simulation to a defined state after an interrupt or solver failure.
I don't see how your property changes are being reversed, unless you are clicking Home->Undo, instead of Home->Previous.
If you are looking for an objective measure of differences between your simulations I suggest you create a plot or monitor (e.g., drag monitor).
If I have already saved a run
If I have already saved a run simulation and cannot get Caedium to run the solver from zero again, will it finish running the simulation?
Is there a way to force the solver to start from zero if the start button is "grayed out?"
After your last reply I was ably to start from zero with my simulation and use it and modify it just as you described for several days.
However, today, when I pull up my simulations and try to run them and modify them, I am again having the same problems I initially described. I have even deleted the residual monitor and replaced it several times, along with doing things to the volume to force mesh regeneration, such as altering the accuracy, etc., but I cannot get the solver to run again or show any activity on the residual monitor.
Any ideas what I might be doing wrong? Can Caedium files become corrupted by users such that they do this? Thanks very much.
Share your project file to diagnose problem
The only way I can think that might cause a project file corruption is to open it using Caedium without the Professional add-on activated and then save it.
The residual monitor plays no part in the simulation controls, it just monitors the simulation activity. Deleting it will make no difference to the simulation control state.
Could it be that you are re-assigning the substance each time you make a change? Once assigned to a volume the substance should be edited by selecting the volume and clicking on the volume tab (far left) in the Properties Panel - the same select-edit procedure applies to all your boundary conditions on faces too.
Also are you aware of the difference between pausing a simulation - toggle the Home->Play button - and interrupting a simulation - toggle the Home->Stop button - which requires an additional toggle to return it to a valid state?
If the Home->Start button is disabled then that simulation is already at its initial position ready to run.
If you still consider this is an error then please share your Caedium project file (zip it first) using Dropbox (or similar) and provide the download link using our Contact form so I can try to diagnose the problem.
Too many streamlines
After assessing your project files I believe the long delays you are seeing when you make changes and the slow project file load times are due to an excessive number of streamlines.
Releasing particles (streamlines) from a face with a fine mesh will create a large number of streamlines which can take a long time to compute and update. I prefer to use streamlines at the end of simulation to avoid the update delays. Also I often create a special seed face which is separate from the simulation so I can control the number of streamlines explicitly using the Accuracy tool rather than having to rely on the simulation mesh density. For an example of this approach see the tutorial "Incompressible Flow Through a Pipe into a Box".
As a much cheaper alternative to streamlines use Color Maps or Arrows on faces - just while the simulation is running. If you add streamlines and want to make a simulation changes then first delete the streamlines and then make the change to avoid slow updates.
I think this is an Apple issue > solution
This issue continued to plague me but I seem to have found the answer.
Invariably, while often being unable to run working simulations elsewhere, they always run when my laptop is at home. I even tried creating new simulations from scratch and running those and the solver would "hang" if I was not at home. The only difference being the location, I figured that it had something to do with whatever wireless network I was using, of all things. I use my home wireless network at home, and that has always worked. I have no idea why it is an issue at other times and over other networks. Over the last several days I have had my laptop with me at work and I simply turned off the wireless receiver. I was subsequently able to run a simulation without it "hanging" at "run volume solver." I tried it several times with the wireless turned on, too, and every time it would "hang." Each time this happened, I then turned the wireless receiver off, and was subsequently able to run the simulation. This may be of help for anyone else who experiences such trouble.
I should say that it appears
I should say that it appears to "hang" because the status bar in the lower right of the window shows a moving blue/white bar that never changes or shows progress.
Sorry. Meant to post this in
Sorry. Meant to post this in the problem section.