|
[Sponsors] |
November 24, 2000, 13:20 |
RCON
|
#1 |
Guest
Posts: n/a
|
I have been working with an irregular structure that requires two sectiones of itself to slide across each other. Naturally I am using a sliding interface (Arbitrary Sliding Interface Method). However, an error popped up during running of the model, noting that error 240 had occurred (miss-matching cell faces), and that the tolerance set in RCON 12 was too low.
The solution was just to set RCON 12 higher ( RCON,12,<new higher value> ). However, on trying to look up info on RCON, i gather that these values are meant to be ignored, untill CD informs us to change them. But my query is, does it matter if the value is set too high, being a tolerance, will a high RCON 12 be detremental at higher values? |
|
November 24, 2000, 13:38 |
Re: RCON
|
#2 |
Guest
Posts: n/a
|
RCON 12 is an area mismatch term ranging from 0 to 1. Star is saying that after it moved the mesh, the new attachments do not line up as well as they used to. A value of less than 0.5 is reasonable, If you need to set it higher, check the following..
1.The two matching faces are on the EXACT same plane 2.The mesh sizes on both sides of the face are close to equal The lower the RCON value, the more accurate the solution at the interface. |
|
November 24, 2000, 14:33 |
Re: RCON
|
#3 |
Guest
Posts: n/a
|
Thanks Pat for the reply,
The execution runs for a couple of steps, and then the error occurs, so what you say about the attachments not aligning up as they used is great. Also for my fist run, it adviced me to change RCON 12 to 0.04, then to 0.08, I ran it eventually at 0.1, but have also already set a model with RCON 12 at 10, just to see the difference. What you say about the range from 0 to 1 is interesting, because looking at my run window, STAR has accepted the tolerance of 10, and appears to be running it fine, but I'll wait to see the final result. With reference to point 1, the answer is yes. But point 2 the answer is no, in one plane. I was hoping to show a picture, but you could probably imagine the two flat planes, sliding across each other, each with irregular grids. Naturally, I will now be running my models with the lowest possible values of RCON 12. Thank you Pat Neuman |
|
November 24, 2000, 14:52 |
Re: RCON
|
#4 |
Guest
Posts: n/a
|
The area mismatch will be different at every iteration, so it is possible that it ran for a while and then crashed at one particular time.
RCON 12 is the maximum the mismatch can be, so you can set it to 10, but it will never exceed 1. You can see the auctual area mismatch which is reported to the star.info file at the start of each iteration of the solver. |
|
November 27, 2000, 05:34 |
Re: RCON
|
#5 |
Guest
Posts: n/a
|
Hi, Mathew,
Pat explains very well about R12. I am just a little bit concerned about your problem - you mentioned flat plane which normally matches quite well with each other, while R12 is designed to allow user to change tolerance of surface matching, it also tries to warn user of potential problem for example the offsets for match are not set properly (offsets are used when the two regions are not at the same place). So if you need a too high R12, please check your mesh at that time step. |
|
November 27, 2000, 06:38 |
Re: RCON
|
#6 |
Guest
Posts: n/a
|
Set "rcon 4 1" which does a mesh preview run and check your mesh movement.
|
|
December 1, 2000, 09:32 |
Re: RCON
|
#7 |
Guest
Posts: n/a
|
hello all again,
thanks Luo for the warning, so i have checked the model at the start, and the vertices match perfectly, the motion of the mesh ensures that the plane of the interface is always in contact with each other. I have run the model well at rcon = 0.1 and thanks to Pat I feel happy with this value. The data I have been getting does make more sense now, infact it looks so wonderful i am very happy. But i do not fully understand what Joern mentioned, I run Star outside Pro, and am unsure about the mesh preview you mentioned. I used Pro to load up my pstt (post transient data) file, and compared the mesh motion, which appears to be as desired. |
|
December 1, 2000, 11:09 |
Re: RCON
|
#8 |
Guest
Posts: n/a
|
It means that you run star without solving for the flow field but only with the mesh motion. You can store the run in the pstt file and check if all works.
|
|
|
|