Hey i am currently trying to get into Molflow+. One thing i am currently running into is that molflow+ interpretes my .stl files strange (exported via Inventor 2018).
The problem is that Molflow+ sees only the solid part of my body, not the “cavity” in between as in your example (https://www.youtube.com/watch?v=anJ21SjEsTw&t=4s). This results in the simulated particles running through the solid part of my vacuum chamber instead through its vacuum volume.
Any idea on how to fix that?
Another question would be how to implement for example windows with a given leakage rate. Intuitively I defined them as “outgassing”, is that right?
Molflow only interprets walls. Therefore the best practice is to model only the vacuum part of the body (i.e. the volume inverse).
Molflow only catches particles from a facet’s front (normal vector side), and is transparent from the back. In an STL file, normal vectors are pointing inside.
Therefore, if you want to use the original CAD model (but be aware that it uses 2x the required facets since it includes the outer shell), select all facets and swap normal (CTRL+N or Facet/Swap normal).
Finally, yes, leakage rates can be simulated with outgassing facets.
I have another question regarding time-dependent simulations.
Is it necessary to set every outgassing/pumping time-dependent when doing some moment based simulations?
It seems to me that only the defined time-dependent outgassing is taken into account, but the priorly defined time-independent properties wont contribute… Am I doing something wrong here?
It is not necessary to have time-dependent parameters for a time-dependent simulation.
When entering a regular number to opacity, sticking or outgassing fields, that number will be taken as a constant parameter
When entering a parameter name to the above boxes, if that parameter is defined in the parameter editor, the corresponding time-dependent value will be taken. If the parameter doesn’t exist, an error message will be shown when starting the simulation.
You cannot have a constant AND a time-dependent outgassing at the same time from the same surface (either-or). I hope I answered the question.
So I have a couple of problems when I try to follow your tutorial video of time dependent simulations:
creating a moment list, it seems you can define the window size the same size as your values (e.g. 1s value, 1s window). This gives me an error of “overlapping time frames”
when I apply a time-dependent sticking factor, I use the calculated value which I get when applying 4l/s pumping speed. Unfortunately, the vice versa calculated pumping speed when defining my time dependent sticking factor (0.113 something for all moments) is 2.28e-310?
I guess that was also the reason why I did not see any pumping in my time-dependent simulation. Is it possible to define a time-dependent pumping speed instead of sticking factor?
If you want to keep the forum “clean” of too much standard questions, I can also contact you via mail if that is preferred. I’m very sorry for the questions, but some things seem non-trivial using Molflow+
Hello,
non overlapping time frames were a compromise to speed-up time-dependent simulations.
The error tells you, that such an overlap exists, e.g.:
0.0,0.1,1.0 window=0.1
1.0,0.1,2.0 window=0.1
will create an overlap, because the last time-bin in line 1 is 1.0+0.5*window_size=1.05 and the first entry of line 2 is 1.0-0.5*window_size=0.95
You will have to set your intervals accordingly.
In case of using a parameter, you can only assign it to the “sticking factor” field, not the pumping speed, at least until the latter is implemented.
When using time-dependent sticking (parameter name written in the sticking field), the pumping speed (an unitialized C++ value, in your case 2.3e-310) should be ignored. It is indeed confusing, so I will hide it in the next version.
I now adapted my timeframes and ignored the pumping speed.
But now I ran into another problem, when I start my time-dependent simulation everything seems good before and while my time-dependent outgassing is happening. But the moment after my time-dependent outgassing stops, where I should be left with only my constant window outgassing, the time-dependent hit counter drops to zero on all facets, pumps etc. included.
Any idea on that? I have my moments defined until 50s, the time-dep. outgassing stops at 10s and at 11s I find 0 hits up to to my 50s moment.
Edit:
I now recognized that when I only apply outgassing from my windows (1E-10mbar l/s) the time-dependent sim works fine. But if I now add a Metal Dispenser (approx. 5E-2mbar l/s) the window outgassing switches from finite Desorption values to zero for all times. Any thoughts on that? I guess I have some definition errors in my sim.
Edit2:
If I lower the Dispenser outgassing number to e.g. 5E-7mbar l/s I see again finite desorption values of the windows. Im quite confused now, is this prior case due to the “high pressure” in the chamber when defining a more or less high outgassing dispenser maybe?
Btw, is there an option to start e.g. at moment 2 or so and let the time run instead of forwarding via the time settings?
Without a look at your geometry, I can’t really interpret Windows and Dispenser, but I believe treating them as “point A” and “point B” should do for the moment.
The constant flow is the final state of the system, therefore if you only have time-dep outgassing, its last value is kept at the end, therefore if the timedep. outgassing ends with 0, the constant flow result will also be 0.
The time-dependent facet hit coutners, however, should show non-zero values for the moments during outgassing (even a bit after until all gas is evacuated). For your data above, it is normal to see 0 values from 11 to 50sec, as I guess all gas leaves the system between 10s (outgassing end) and 11s.
I didn’t understand your last question, could you elaborate?
I am sorry, that question was written quite stupid.
Let me rephrase, is there a difference in changing the moments from e.g. 2 to 3 (if 2 is e.g. at 2s and 3 is at 3s) via the moment editor or selecting the 2nd moment and let the simulation time run for in this case 1s to “reach” the 3rd moment.
We maybe misunderstanding each other, but if that was your question: the computation time (how long Molflow runs) has nothing to do with physical time.
All moment results are calculated at the same time. You could (and often should) run Molflow for hours to simulate a process from 0 to 1s. Molflow doesn’t simulate time-wise (in chronologic order).
Is this what you asked?