Author |
Message |
Jerry Brown (jerrybrown11743)
Member Username: jerrybrown11743
Post Number: 8 Registered: 03-2004
| Posted on Wednesday, May 05, 2004 - 04:52 pm: | |
I've been searching for a way to make a two-stage calculation in which I feedback some of the results of the first stage to redefine one of the domain boundaries for the second stage. You're suggestion of using ARRAY(#include "yfile" 0) looked promising. However, I can't seem to control the number of data points when I build "yfile" and that makes it impossible to use the REPEAT construct to build the boundary. If ARRAY would accept arguments using the VAL function, I could get the things to work. But, it won't. Also, I can't use variables in the boundary definition even though they are coming from a table in the DEFINITIONS section. Attached is an example of what I wish I could do. The pertinent parts of the program are preceded by comments that start in the first column. |
Patricio A. Greco (pagreco)
Member Username: pagreco
Post Number: 24 Registered: 08-2003
| Posted on Friday, May 07, 2004 - 08:53 am: | |
There aren't possibility to redefine the boundary specification in FLEXPDE , but you can do it runing the program many times as Batchall.pde |
Jerry Brown (jerrybrown11743)
Member Username: jerrybrown11743
Post Number: 9 Registered: 03-2004
| Posted on Friday, May 07, 2004 - 11:57 am: | |
I understand that the boundary definitions can't be changed while the problem is running. My plan was to use the staging capability to run the problem in the usual way the first time and write the result to a file. Then, on the second run the domain specifications would be developed in the "Definitions" section based on the values in the file from the first run. This is almost possible except for the problems mentioned above. |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 147 Registered: 06-2003
| Posted on Friday, May 07, 2004 - 01:26 pm: | |
We seem to have you blocked at all turns. Attached are two files that implement a kludge to do what you want. I seem to have broken something, though, because it doesn't give the same solution as the original. f1.pde is supposed to be your original file, with a SUMMARY output that writes each of the 21 moved points explicitly. f2.pde reads this file in and sets the boundary accordingly. This seems to be the best I can do until we get our moving boundary version working.
|