Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
post-processing: force evaluation at point and edges.
Posted 7 avr. 2011, 12:04 UTC−4 Results & Visualization, Structural Mechanics Version 4.1 10 Replies
Please login with a confirmed email address before reporting spam
Hello everyone,
The problem I faced is following. I have a simple model of the tower with plate mounted on the top of it. Tower is fixed-bottom (fixed constrained). Tower is subjected to periodic loads on one edge (defined as total force). Besides the constant gravitational force applied as body loads (force per unit volume, solid.rho*gravitational constant).
After running my model I try to extract an information about forces generated at the place where plate contacts the tower. I am interesting on both, point forces and forces on edges. I could not use reaction force option because it's only applicable to constraint boundary.
So is there a way to get such an information?
I attached a model and would appreciate any help.
Best regards,
Artem
The problem I faced is following. I have a simple model of the tower with plate mounted on the top of it. Tower is fixed-bottom (fixed constrained). Tower is subjected to periodic loads on one edge (defined as total force). Besides the constant gravitational force applied as body loads (force per unit volume, solid.rho*gravitational constant).
After running my model I try to extract an information about forces generated at the place where plate contacts the tower. I am interesting on both, point forces and forces on edges. I could not use reaction force option because it's only applicable to constraint boundary.
So is there a way to get such an information?
I attached a model and would appreciate any help.
Best regards,
Artem
Attachments:
10 Replies Last Post 11 avr. 2011, 01:30 UTC−4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
a few comments on your model
1) you have NOT defined where the loads apply, you must select the respective "entities"
2) one should avoid edge and point loads because this produces singularities but if you are not interested in the stress, you can still get useful info by using them
3) if you want to see the frequency response, use the frequency domain solver instead of a transient with a sinus load
4) in your simple case you can use "assembly mode" and set an identity boundary condition where the two parts match, and turn on the weak constraints, and look at these for the loading, but I would suggest to integrate over an area instead of using "points or edges"
--
Good luck
Ivar
a few comments on your model
1) you have NOT defined where the loads apply, you must select the respective "entities"
2) one should avoid edge and point loads because this produces singularities but if you are not interested in the stress, you can still get useful info by using them
3) if you want to see the frequency response, use the frequency domain solver instead of a transient with a sinus load
4) in your simple case you can use "assembly mode" and set an identity boundary condition where the two parts match, and turn on the weak constraints, and look at these for the loading, but I would suggest to integrate over an area instead of using "points or edges"
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thank you, Ivar.
I specified an "entities" where the loads applied.
I will go for your suggestion about creating an assembly.
Should I define weak constraint on the Identity boundary pair?
Besides, as far as I understand, after creating an assembly and defining an Identity pair, I have to go to Solid Mechanics and defined the boundary condition (e.g. Continuity or rigid connector), right?
Also, when you said, integrate over the area, you meant for post-processing or as input loads?
You may look for attached model if you need.
Best regards,
Artem Korobenko
I specified an "entities" where the loads applied.
I will go for your suggestion about creating an assembly.
Should I define weak constraint on the Identity boundary pair?
Besides, as far as I understand, after creating an assembly and defining an Identity pair, I have to go to Solid Mechanics and defined the boundary condition (e.g. Continuity or rigid connector), right?
Also, when you said, integrate over the area, you meant for post-processing or as input loads?
You may look for attached model if you need.
Best regards,
Artem Korobenko
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
if you go for an assembly of two parts, then you can apply/check the weak form to the identity continuity pair and analyse the "lm" values to extract/postprocess the most precise forces.
Use the continuity pair BC (I would not necessarily use the RBC (rigid boundary connector) as it stiffens the domain you select, or perhaps to define the boundary load, then its easier to apply a pure moment)
again to identify the forces, it's always better/more precise to integrate over an area of several mesh elements than selecting just a point hence a node, or an edge, as on edges and points the forces are defined as averages from the adjacent boundaries, or edges respectively
--
Good luck
Ivar
if you go for an assembly of two parts, then you can apply/check the weak form to the identity continuity pair and analyse the "lm" values to extract/postprocess the most precise forces.
Use the continuity pair BC (I would not necessarily use the RBC (rigid boundary connector) as it stiffens the domain you select, or perhaps to define the boundary load, then its easier to apply a pure moment)
again to identify the forces, it's always better/more precise to integrate over an area of several mesh elements than selecting just a point hence a node, or an edge, as on edges and points the forces are defined as averages from the adjacent boundaries, or edges respectively
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thank you, Ivar,
It is working, but the only value that I can extract is displacement (Lagrange multiplier for displacement field), and I could not find anything similar for the forces?
Best regards,
Artem
It is working, but the only value that I can extract is displacement (Lagrange multiplier for displacement field), and I could not find anything similar for the forces?
Best regards,
Artem
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
now I'm confused, for me you should get the LM's that once integrated gives you the forces, it's true that the units are not yet implemented on the lm's in v4.1
--
Good luck
Ivar
now I'm confused, for me you should get the LM's that once integrated gives you the forces, it's true that the units are not yet implemented on the lm's in v4.1
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Ivar,
When you said LM....what exact variable you meant?
Because the only one I can use it is e.g. u_lm,v_lm,w_lm.
Could you write the name of the variable that I should integrate over the surface to get the forces. In my mind it should be something like solid.Force_lm.
Also, is it possible to model a contact pair with two curved boundaries? (Because it says that "source boundary" should be flat) ?
Best,
Artem
When you said LM....what exact variable you meant?
Because the only one I can use it is e.g. u_lm,v_lm,w_lm.
Could you write the name of the variable that I should integrate over the surface to get the forces. In my mind it should be something like solid.Force_lm.
Also, is it possible to model a contact pair with two curved boundaries? (Because it says that "source boundary" should be flat) ?
Best,
Artem
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I agree I should have, but I did not have hands on my COMSOL when I answered.
So we are talking about the same u_lm, v_lm, w_lm in V4 notation
which for me should have units [Pa] see the doc
StructuralMechanicsModuleUsersGuide.pdf p25++ (v4.1)
--
Good luck
Ivar
I agree I should have, but I did not have hands on my COMSOL when I answered.
So we are talking about the same u_lm, v_lm, w_lm in V4 notation
which for me should have units [Pa] see the doc
StructuralMechanicsModuleUsersGuide.pdf p25++ (v4.1)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Artem,
Regarding your contact question, there is no restriction on either contact surfaces (source and destination) being flat. Both can be curved.
Nagi Elabbasi
Veryst Engineering
Regarding your contact question, there is no restriction on either contact surfaces (source and destination) being flat. Both can be curved.
Nagi Elabbasi
Veryst Engineering
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thank you, Ivar and Nagi,
One more question. I used the approach with defining identity pair, with weak constraint. Then, I integrate over the surface e.g. w_lm. Then I multiply this value by surface area and receive force in z direction, am I write? Can I somehow check this value by force balance on this boundary?
What about, if instead of Identity pair I have contact pair, should I use the same way to define forces over not constrained surface? (For constrained I could simply use Reaction force option from post-processing me
Also, if I use weak constraint for Identity pair, does it mean that I should use them also for fixed constraint on the other boundaries (or I can use the regular one)?
Best regards,
Artem
One more question. I used the approach with defining identity pair, with weak constraint. Then, I integrate over the surface e.g. w_lm. Then I multiply this value by surface area and receive force in z direction, am I write? Can I somehow check this value by force balance on this boundary?
What about, if instead of Identity pair I have contact pair, should I use the same way to define forces over not constrained surface? (For constrained I could simply use Reaction force option from post-processing me
Also, if I use weak constraint for Identity pair, does it mean that I should use them also for fixed constraint on the other boundaries (or I can use the regular one)?
Best regards,
Artem
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
No, you are not fully right, u_lm being in Pa or N/m^2, if you integrate that over an area you are already multiplying it by m^2 from the dx*dy of the integration so it's enough to integrate to get out total forces in Newtons
Contact pairs is another issue, as a contact means continuity if the two parts are pressed one against the other in a normal direction, pressure + sliding forces if the load direction is not fully perpendicular and free if you are pulling. This is far more complex to study and to make converge
--
Good luck
Ivar
No, you are not fully right, u_lm being in Pa or N/m^2, if you integrate that over an area you are already multiplying it by m^2 from the dx*dy of the integration so it's enough to integrate to get out total forces in Newtons
Contact pairs is another issue, as a contact means continuity if the two parts are pressed one against the other in a normal direction, pressure + sliding forces if the load direction is not fully perpendicular and free if you are pulling. This is far more complex to study and to make converge
--
Good luck
Ivar