|
[Sponsors] |
March 18, 2005, 06:01 |
Iterations
|
#1 |
Guest
Posts: n/a
|
I'm using STAR CD version 3.24. If I set the number of iterations at 1000000 or more,
? FORTRAN Runtime Error: ? Illegal character in numeric imput ? READ(UNIT=10,... I don't need 1000000 iterations, it was just a large number to ensure that it wouldn't stop before meeting the residual conditions, but I wasted a lot of time figuring out that this was the problem rather than the rest of my setup. AFAIK there is no error message other than the above. Has anyone else had this problem? Is there a way to ensure it doesn't happen? |
|
March 18, 2005, 15:59 |
Re: Iterations
|
#2 |
Guest
Posts: n/a
|
The easy solution is don't set the number of iterations to 1000000. If you look at the problem file you'll see that where the iteration number is, there is just *********.
As a rule of thumb, think about what you are doing, and don't enter stupid numbers into CFD codes. |
|
March 19, 2005, 05:33 |
Re: Iterations
|
#3 |
Guest
Posts: n/a
|
Of course I realise not to use 1000000, but this is not the point. THe only stupidity on my part was not checking the problem file. But there is nothing wrong with entering an arbitrarily large number to ensure that it will not stop unless it has converged. My point was only to note that the message STAR gives you is not easy to relate to the cause.
And perhaps since I've only been using STAR since January, there is a possibility I might enter stupid numbers from time to time..... we all start somewhere |
|
March 21, 2005, 13:10 |
Re: Iterations
|
#4 |
Guest
Posts: n/a
|
You didn't make a stupid mistake and in fact the error message didn't even come from STAR - it came from the OS or more probably the run-time component of the compiler as it tried to read an invalid field. In a perfect world, when you entered the number, pro-STAR should have prevented you from entering something that big if it couldn't write it out. On the other hand, any given analysis requires thousands of numbers to be entered and checked for sanity and occasionally one slips by the checks because the programmer never considered the possibility or for some other disconnect. You got caught in an unlucky combination. If you report it to the support people it should get fixed so it doesn't happen to you or the next guy again.
|
|
March 21, 2005, 16:46 |
Re: Iterations
|
#5 |
Guest
Posts: n/a
|
That's a good idea, thanks, I'l do that
|
|
March 24, 2005, 16:30 |
Re: Iterations
|
#6 |
Guest
Posts: n/a
|
Not sure I agree. There is everything wrong with entering arbirtraily large numbers, especially in a Fortran program. Even Excel will throw a wobbler if you overflow it.
Part of being a good CFD engineer is not leaving these things to chance. Every value you enter should be considered. Bad workmen don't blame their tools |
|
March 25, 2005, 13:50 |
Re: Iterations
|
#7 |
Guest
Posts: n/a
|
"Bad workmen don't blame their tools"
Whom do they blame then? Well, I admit it was a bad idea and it was my mistake, but I don't think it would be unreasonable for it to be documented for the beginners, or at least throw up an error message that can be connected.. but this is perhaps an obvious point for somebody with experience and there are probably other things they need to work on. Thanks for making me think about the problem a bit more though Ant |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
UDF of iterations | h6h6 | FLUENT | 4 | August 28, 2018 23:07 |
Iterations | Jorge A | FLUENT | 0 | September 6, 2007 13:24 |
iterations | suailek | FLUENT | 0 | July 30, 2004 05:56 |
Sub-iterations - What do they mean? | Arun | FLUENT | 0 | May 18, 2004 14:37 |
Iterations + more | Steve | Main CFD Forum | 1 | March 23, 2004 12:18 |