Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Defining a probe relative to a moving mesh boundary

Please login with a confirmed email address before reporting spam

I am working with a model which is using the moving mesh interface to oscillate a domain(plate) in my model. I know how to define probes based on static coordinates but I need to define a probe relative to the position of my oscillating plate. For example in the attached model I would like to define a point probe which is always 0.05m away from boundary 5. Since boundary 5 is oscillating I need to be able to define this probes position relative to boundary 5.

Under Model1 - Definitions - Domain Point Probe 1 - Point Selection is it possible do to something like this;

x: 0 y: boundary5 - 0.05?

If I knew the correct reference notation for "boundary5" would this work?


13 Replies Last Post Sep 10, 2010, 2:36 p.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 30, 2010, 5:34 p.m. EDT
Ditto! I couldn't find a way too, since it does not a have a spatial option.
Ditto! I couldn't find a way too, since it does not a have a spatial option.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 30, 2010, 5:41 p.m. EDT
Ok, that didn't take long :)

- Define an integral probe:
cl.ly/b730da4b4d8dfab72a15

- Define a variable
cl.ly/f9c5f7ab754e7786feee

- Use a global probe:
cl.ly/b69d397f2db2b6ba8707

- You get this:
cl.ly/ff4a99042fbbb5c05293

Right? ;)
Ok, that didn't take long :) - Define an integral probe: http://cl.ly/b730da4b4d8dfab72a15 - Define a variable http://cl.ly/f9c5f7ab754e7786feee - Use a global probe: http://cl.ly/b69d397f2db2b6ba8707 - You get this: http://cl.ly/ff4a99042fbbb5c05293 Right? ;)

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 31, 2010, 3:48 a.m. EDT
Hi Danial

Do you ot belieeve its easier like this fo keep trackon the long term ?

Still thanks for the procedure:)
--
Good luck
Ivar
Hi Danial Do you ot belieeve its easier like this fo keep trackon the long term ? Still thanks for the procedure:) -- Good luck Ivar


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 31, 2010, 5:41 a.m. EDT

Hi Danial

Do you ot belieeve its easier like this fo keep trackon the long term ?

Still thanks for the procedure:)
--
Good luck
Ivar


Right. It was just easier with cloud app as it takes screenshot, uploads, and make a short link into clipboard already with a quick shortcut. lazy guy here :)

I will upload to the forum next time.

Danial
[QUOTE] Hi Danial Do you ot belieeve its easier like this fo keep trackon the long term ? Still thanks for the procedure:) -- Good luck Ivar [/QUOTE] Right. It was just easier with cloud app as it takes screenshot, uploads, and make a short link into clipboard already with a quick shortcut. lazy guy here :) I will upload to the forum next time. Danial

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 31, 2010, 7:16 a.m. EDT
Hi

thanks,
it basically because it's more reliable on long term, I have found good replies from 6 months agoo, but the links are then dead.

Stay somewhat lazy, it's often what save your life ;)

--
Good luck
Ivar
Hi thanks, it basically because it's more reliable on long term, I have found good replies from 6 months agoo, but the links are then dead. Stay somewhat lazy, it's often what save your life ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 31, 2010, 2:51 p.m. EDT
Thanks for the information. I think I am still missing the link between the model coupling and being able to make a probe that plots more than just the "y" position.

For example in the attached model above, I had a domain point probe which was offset 0.05m from the bottom of boundary 5(on my oscillating plate). When the plate starts to oscillate the probe remains exactly in it's original position(x:0 y:-0.05). What I need to do instead is have the probe move with the oscillating plate but always be at a constant offset position. The formula for the oscillating plate is 0.1*sin(2*pi*2*t). So at t=0.125s, the plate has moved postive 0.1m in the y direction therefore I need my probe to be at x:0, y:0.05 at this time.

I added an offset to my posy variable, see posy_offset.png. Now when doing a Global Variable Probe I get a plot of the offset position from the oscillating plate, i.e. a sinusoid from -0.05 to 0.05. This is great but what I really need is to be able to plot the other model parameters like Voltage and Electric field at this offset oscillating position. I then went back to my Domain Point Probe and put in "posy" for my y position, see domain_no_posy.png. However as can be seen in the screenshot I get a unknown model parameter error when entering it. COMSOL seems to ignore this change and use the previously defined static postion. Do you have any tips?


Thanks for the information. I think I am still missing the link between the model coupling and being able to make a probe that plots more than just the "y" position. For example in the attached model above, I had a domain point probe which was offset 0.05m from the bottom of boundary 5(on my oscillating plate). When the plate starts to oscillate the probe remains exactly in it's original position(x:0 y:-0.05). What I need to do instead is have the probe move with the oscillating plate but always be at a constant offset position. The formula for the oscillating plate is 0.1*sin(2*pi*2*t). So at t=0.125s, the plate has moved postive 0.1m in the y direction therefore I need my probe to be at x:0, y:0.05 at this time. I added an offset to my posy variable, see posy_offset.png. Now when doing a Global Variable Probe I get a plot of the offset position from the oscillating plate, i.e. a sinusoid from -0.05 to 0.05. This is great but what I really need is to be able to plot the other model parameters like Voltage and Electric field at this offset oscillating position. I then went back to my Domain Point Probe and put in "posy" for my y position, see domain_no_posy.png. However as can be seen in the screenshot I get a unknown model parameter error when entering it. COMSOL seems to ignore this change and use the previously defined static postion. Do you have any tips?


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 31, 2010, 6:24 p.m. EDT
Hi

I mostly split my object with a line/surface (interiour continuous boundary) and use that to define the probe plot line/area, if I understand you well this could be one way around.

Perhaps by playing with the other reference variables X,Y or Xm Ym you might find a way too, but the one above is probably the simplest one

--
Good luck
Ivar
Hi I mostly split my object with a line/surface (interiour continuous boundary) and use that to define the probe plot line/area, if I understand you well this could be one way around. Perhaps by playing with the other reference variables X,Y or Xm Ym you might find a way too, but the one above is probably the simplest one -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 7, 2010, 2:12 p.m. EDT
Thank you for your reply. I'm not clear on how splitting my object could help though. Could you provide some more information regarding that?

One issue I can't seem to work around is the fact that a Domain Point Probe seems to require a number entered into it's coordinates. I can not enter an expression or a variable. Therefore it doesn't seem like I can create a Domain Point Probe which moves with time.
Thank you for your reply. I'm not clear on how splitting my object could help though. Could you provide some more information regarding that? One issue I can't seem to work around is the fact that a Domain Point Probe seems to require a number entered into it's coordinates. I can not enter an expression or a variable. Therefore it doesn't seem like I can create a Domain Point Probe which moves with time.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 8, 2010, 1:38 a.m. EDT
Hi

I agree splitting you object will only define a region fixed w.r.t to the geometry, and moving with the geometry. But you have mostly three reference frames in COMSOL, spatial (original), material (typically the deformed shapes in Structural) and mesh (as mesh might be different from the deformation i.e. cfd, ALE ...).

But I do not have any obvious answer to propose, would need some thinking and testing (have you tried to push it to "support" ?). It is though of interest for many cases, I agree

--
Good luck
Ivar
Hi I agree splitting you object will only define a region fixed w.r.t to the geometry, and moving with the geometry. But you have mostly three reference frames in COMSOL, spatial (original), material (typically the deformed shapes in Structural) and mesh (as mesh might be different from the deformation i.e. cfd, ALE ...). But I do not have any obvious answer to propose, would need some thinking and testing (have you tried to push it to "support" ?). It is though of interest for many cases, I agree -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 8, 2010, 3:48 p.m. EDT

Thank you for your reply. I'm not clear on how splitting my object could help though. Could you provide some more information regarding that?

One issue I can't seem to work around is the fact that a Domain Point Probe seems to require a number entered into it's coordinates. I can not enter an expression or a variable. Therefore it doesn't seem like I can create a Domain Point Probe which moves with time.


I posted the only way I found so far. The problem is as you said, unlike v3.5a, Comsol 4 does not accept predefined points and frames but only coordinates. And, this is not helpful for the case of ALE. The only place you can define points and frame is coupling. So the way I posted is a workaround. If you find a better way, please post here.
[QUOTE] Thank you for your reply. I'm not clear on how splitting my object could help though. Could you provide some more information regarding that? One issue I can't seem to work around is the fact that a Domain Point Probe seems to require a number entered into it's coordinates. I can not enter an expression or a variable. Therefore it doesn't seem like I can create a Domain Point Probe which moves with time. [/QUOTE] I posted the only way I found so far. The problem is as you said, unlike v3.5a, Comsol 4 does not accept predefined points and frames but only coordinates. And, this is not helpful for the case of ALE. The only place you can define points and frame is coupling. So the way I posted is a workaround. If you find a better way, please post here.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 9, 2010, 5:37 a.m. EDT
ok :) I see your point now.

What you should integrate is intop1(y-Y), which is the ALE displacement.
ok :) I see your point now. What you should integrate is intop1(y-Y), which is the ALE displacement.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 9, 2010, 4:42 p.m. EDT
Thank you again for thinking about this issue. I'm still not sure exactly how to get to the Voltage or Electric Field. Creating the posy variable as outlined allows one to track the oscillation of the moving mesh(or an offset from it) with time. I now need to figure out how to determine the Voltage or Electric Field at the positions given by posy Vs time. For Example I'd like to know the following;

posy - is given by intop1(y)-0.05 based on the above method. using intop1(y-Y) returned the same result. see the attachment for a plot of posy Vs time. This represents the point(in spatial y domain) that is offset -0.05 from the bottom of the moving mesh(point 3), controlled by 0.1*sin(2*pi*2*t).

t, posy, es.Ey(posy), V(posy)
0, -0.05, ?, ?
0.125, 0.05, ?,?
0.375, -0.15, ?, ?

where es.Ey(posy) means the Electric Field in the Y direction at location x = 0, y = posy. Similarly for V(posy). I don't see how I can use a Global Variable Probe to get the Electric Field or Voltage though. Is this possible? When I try to use a Domain Point Probe with coordinates x = 0, y = posy, I get the uknown model parameter error mentioned about. I am somewhat new to COMSOL so if there is a way to get the Electric Field and Voltage at this varying(w/r/t time) y position let me know. I will probably contact support tomorrow regarding this. Thanks again!
Thank you again for thinking about this issue. I'm still not sure exactly how to get to the Voltage or Electric Field. Creating the posy variable as outlined allows one to track the oscillation of the moving mesh(or an offset from it) with time. I now need to figure out how to determine the Voltage or Electric Field at the positions given by posy Vs time. For Example I'd like to know the following; posy - is given by intop1(y)-0.05 based on the above method. using intop1(y-Y) returned the same result. see the attachment for a plot of posy Vs time. This represents the point(in spatial y domain) that is offset -0.05 from the bottom of the moving mesh(point 3), controlled by 0.1*sin(2*pi*2*t). t, posy, es.Ey(posy), V(posy) 0, -0.05, ?, ? 0.125, 0.05, ?,? 0.375, -0.15, ?, ? where es.Ey(posy) means the Electric Field in the Y direction at location x = 0, y = posy. Similarly for V(posy). I don't see how I can use a Global Variable Probe to get the Electric Field or Voltage though. Is this possible? When I try to use a Domain Point Probe with coordinates x = 0, y = posy, I get the uknown model parameter error mentioned about. I am somewhat new to COMSOL so if there is a way to get the Electric Field and Voltage at this varying(w/r/t time) y position let me know. I will probably contact support tomorrow regarding this. Thanks again!

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 10, 2010, 2:36 p.m. EDT
Since the attachment didn't work yesterday here it is. I will contact support regarding this issue and will post any new information I recieve.
Since the attachment didn't work yesterday here it is. I will contact support regarding this issue and will post any new information I recieve.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.