Discussion Forum

Run the comsol installation DVD and point to the correct matlab directory. Then it should be able to find matlab.

OK there are two alternative ways to solve this Matlab connection .

begin
if (Matlab is installed in your local machine) then

if (you are using MS-win)then
go to your comsol\bin folder and look for the file comsol.opts
open this script with a text editor and look for the variable MLROOT
it should be in the beginning of the file.
put to this variable the path of the matlab folder
ex: MLROOT=blabla\blabla\matlab

elseif (you are using Linux OS system) then
go to your comsol/bin/comsol.sh
open this script with a text editor and again look for the variable MLROOT
it should be in the beginning of the file.
put to this variable the path of the matlab folder
ex: MLROOT:=blabla/blabla/matlab

endif

else
re-install your comsol by giving the path of the matlab folder during the installaion.
endif
end

I hope it helps

Cheers

Thanks for the hint!

COMSOL was looking for something to which it might append \bin\win32
so I added
mlroot=\Program Files\MATLAB\R2009a\

which launched MATLAB.

MATLAB then generated an error:
??? Undefined function or method 'comsolmexpath' for input arguments of type 'char'.

The saga continues...

~Rich~

Hi

once you have the matlab root OK I use:

>> comsol matlab -np 1 -tmpdir /export/home/scratch/$USER

to start "matlab" with the comsol library link loaed and OK, then I type "comsol" in the matlab environment to get my COMSOL GUI, then back and forth wth CNTRL-F and "file import ..."


the "comsol matlab" command is the minimum "-np 1" is to use 1 core, or replace by 8 if you have so many, the "-tempdir ..." is to force comsol to use the local disk for it's scratch file and not to scatter them on others, that in my case are network server directoris.

This is expalined in the installation guide.

Good luck
Ivar

OK
So you have MS-win as OS !!
The path that you have assigned is wrong. This is not the Matlab folder.
Depending on the chosen folder during the Matlab installation your Matlab folder should be in C:\Matlab.. but not
"\Program Files\MATLAB\R2009a\" This is wrong.

Cheers

Let me try to reply to all the notes, because I am not sure that I can reply to a specific one.

I have 3 MATLAB versions installed and they are in:
C:\Program Files\MATLAB\R2007a
C:\Program Files\MATLAB\R2008a
C:\Program Files\MATLAB\R2009a
...and they all seem to work OK. I know MATLAB has some peculiar naming conventions, but the installations seem to work OK.

I keep my data in c:\MATLAB\folder...

I removed COMSOL completely and reinstalled it. I pointed it to MATLAB 2007a because the manual I have says that COMSOL won't work with anything later than that. Maybe it will, but the 2007a version is also OK, it seems.

I am able to call MATLAB from the COMSOL UI
File>Client/Server/MALAB>MATLAB
and it works.

I am now trying to accomplish my main goal, that of tracking many particles. So far, I have mixed results. I wrapped a loop around

for i = 0:2
part_loc_x = [0.6+(i*0.02)];
part_loc_y = [0.6+(i*0.02)];
part_loc_z = [0.6+(i*0.02)];
postplot(fem, ...

My first attempt generated the whole plot each time, but only the last particle plotted. So then I tried to execute the original
postplot(fem, ...
followed by a loop to plot just my particles, but I have not found how to do this yet. It seems that I must study each parameter in postplot() and see which ones to use. I would appreciate any feedback here, too.

Many thanks to all of you for your time.
~R~

p.s.

Fortunately, we can edit our posts here. I settled on this to present:
hold on;
for i = 1:5
part_loc_x = [0.6+(rand*0.1)];
part_loc_y = [-0.5+(rand*1.0)];
part_loc_z = [-0.5+(rand*1.0)];
% [part_loc_x part_loc_y part_loc_z]
partvelstart_x = num2str(1e5*rand, '%10.5e\n');
partvelstart_y = num2str(1e5*rand, '%10.5e\n');
partvelstart_z = num2str(1e5*rand, '%10.5e\n');
postplot(fem, ...
'flowdata',{'Ex_es','Ey_es','Ez_es'}, ...
'flowcolor',[1.0,0.0,0.0], ...
'flowlines',30, ...
'flowtype','line', ...
'flowtol',0.01, ...
'flowstattol',0.01, ...
'flowlooptol',0.01, ...
'flowmaxtime',Inf, ...
'flowmaxsteps',5000, ...
'flowback','on', ...
'flownormal','off', ...
'partmasstype','mass', ...
'partplotas','lines', ...
'parttstart','auto', ...
'parttype','line', ...
'partdroptimes','auto', ...
'partstart',{part_loc_x,part_loc_y,part_loc_z}, ...
'partlinecolor',[0.0,0.8,0.8], ...
'partdata',{'-partq*Ex_es','-partq*Ey_es','-partq*Ez_es'}, ...
'partvelstart',{partvelstart_x,partvelstart_y,partvelstart_z}, ...
'partmass','9.1095e-31', ...
'const',{'partq','1.602e-19'}, ...
'partrtol',0.001, ...
'partatol','auto', ...
'parthstart','auto', ...
'parthmax','auto', ...
'partstatictend','auto', ...
'partmaxsteps','auto', ...
'partedgetol',0.001, ...
'partvelvar',{'partu','partv','partw'}, ...
'parttvar','partt', ...
'partstatic','off', ...
'partres',5, ...
'solnum','end', ...
'phase',0, ...
'refine','auto', ...
'geom','on', ...
'geomnum',1, ...
'bdl',{[3,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22]}, ...
'sdl',{[1]}, ...
'transparency',1.0);
end
hold off;

It may well be that I can eliminate more of these parameters and still get just the plots of the particles.

~R~