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 The information provided may be out of date.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.
Problems related to relative tolerance and relative error
Posted 24 févr. 2013, 13:36 UTC−5 Structural Mechanics Version 4.3a 5 Replies
Please login with a confirmed email address before reporting spam
Hello
I want to reproduce a model in which there are two spheres in contact one upon the other along the z axis, and then i has applied two forces, one on the top (point or boundary region of the sphere above) with -z direction and other force on the bottom (point or boundary region of the sphere below) with z direction. In summary the idea is to confront the two spheres subject for the forces and will see the results corresponding to the stress and displacement of both spheres. Unfortunately when I solve and compute it, appear the following messages:
Failed to find a solution.
The relative residual (55) is greater than the relative tolerance.
Returned solution is not converged.
Failed to find a solution.
The relative error (2.2e2) is greater than the relative tolerance.
Returned solution is not converged.
What is the meaning of this? the reason is because the problem isn't configure correctly and therefore cannot converged? How can I avoid this problem and resolve it succesfully?
Thanks beforehand.
I want to reproduce a model in which there are two spheres in contact one upon the other along the z axis, and then i has applied two forces, one on the top (point or boundary region of the sphere above) with -z direction and other force on the bottom (point or boundary region of the sphere below) with z direction. In summary the idea is to confront the two spheres subject for the forces and will see the results corresponding to the stress and displacement of both spheres. Unfortunately when I solve and compute it, appear the following messages:
Failed to find a solution.
The relative residual (55) is greater than the relative tolerance.
Returned solution is not converged.
Failed to find a solution.
The relative error (2.2e2) is greater than the relative tolerance.
Returned solution is not converged.
What is the meaning of this? the reason is because the problem isn't configure correctly and therefore cannot converged? How can I avoid this problem and resolve it succesfully?
Thanks beforehand.
5 Replies Last Post 11 mars 2013, 12:24 UTC−4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
contact issues are delicate, and using point forces is also very delicateas you create singularities, its better to use a pressure region (boundary) of finite size, or to consider a half or even better quarter sphere with one resp. two symmetry planes
--
Good luck
Ivar
contact issues are delicate, and using point forces is also very delicateas you create singularities, its better to use a pressure region (boundary) of finite size, or to consider a half or even better quarter sphere with one resp. two symmetry planes
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thanks a lot Ivar.
Then there are no converged solutions for my model that I has explained you before. To introduce the pressure I only need to input P instead F on the parameters? I don't understand what is the implication to put symmetry planes on the model.
I have another question. I want build a semisphere but i want it is in contact upon all the base. This implies which this semisphere must have a base on the bottom on the contrary this semisphere will have only a point contact with the base. How could I build this semisphere?
An alternative for my model that I explain at the first time I have thought to fixed the laterals points of the spheres but maybe is not a good idea.
Also I have seen that I need another module to reproduce the consequence the contact after applying the forces without cross them, instead after apply the forces one sphere penetrates in another. What is the module to use?
Thanking you in advance for your attencion Ivar.
Then there are no converged solutions for my model that I has explained you before. To introduce the pressure I only need to input P instead F on the parameters? I don't understand what is the implication to put symmetry planes on the model.
I have another question. I want build a semisphere but i want it is in contact upon all the base. This implies which this semisphere must have a base on the bottom on the contrary this semisphere will have only a point contact with the base. How could I build this semisphere?
An alternative for my model that I explain at the first time I have thought to fixed the laterals points of the spheres but maybe is not a good idea.
Also I have seen that I need another module to reproduce the consequence the contact after applying the forces without cross them, instead after apply the forces one sphere penetrates in another. What is the module to use?
Thanking you in advance for your attencion Ivar.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
pressure of distributed force are the same, but do not apply them onto a "point". Use rather a "Boundary Load" on a "Boundary - Entity". (domains a re volumes in 3D for surfaces as boundaries, as in 2D a domain is a surface and a boundary is a line. (Points are boundaries in 1D only so there they are OK to be used ;)
Then using a force or a pressure is rather a quation of convinience, the "total force" is just distributing the average force value over the total area (length*depth in 2d) of the boundary. Note that pressures are scalars ALWAYS normal to the boundary, as Forces are vector components.
To simulate contact try a simple model first to lear how to set up the contact physics and to tweak the sovlers
--
Good luck
Ivar
pressure of distributed force are the same, but do not apply them onto a "point". Use rather a "Boundary Load" on a "Boundary - Entity". (domains a re volumes in 3D for surfaces as boundaries, as in 2D a domain is a surface and a boundary is a line. (Points are boundaries in 1D only so there they are OK to be used ;)
Then using a force or a pressure is rather a quation of convinience, the "total force" is just distributing the average force value over the total area (length*depth in 2d) of the boundary. Note that pressures are scalars ALWAYS normal to the boundary, as Forces are vector components.
To simulate contact try a simple model first to lear how to set up the contact physics and to tweak the sovlers
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thanks for answers Ivar.
I don't understand this:
Failed to find a solution.
Segregated group 1
Singular matrix.
There are 356 void equations (empty rows in matrix) for the variable mod1.u.
at coordinates: (7.14788e-09,-2.00713e-08,4.01192e-08), (5.92103e-09,-1.98664e-08,3.72976e-08), (3.01841e-09,-2.10876e-08,4.0093e-08), (1.36218e-08,-1.63453e-08,3.99136e-08), (1.1648e-08,-1.70165e-08,3.69161e-08), ...
There are 356 void equations (empty rows in matrix) for the variable mod1.v.
at coordinates: (7.14788e-09,-2.00713e-08,4.01192e-08), (5.92103e-09,-1.98664e-08,3.72976e-08), (3.01841e-09,-2.10876e-08,4.0093e-08), (1.36218e-08,-1.63453e-08,3.99136e-08), (1.1648e-08,-1.70165e-08,3.69161e-08), ...
There are 449 void equations (empty rows in matrix) for the variable mod1.w.
at coordinates: (7.14788e-09,-2.00713e-08,4.01192e-08), (5.92103e-09,-1.98664e-08,3.72976e-08), (3.01841e-09,-2.10876e-08,4.0093e-08), (1.36218e-08,-1.63453e-08,3.99136e-08), (1.1648e-08,-1.70165e-08,3.69161e-08), ...
and similarly for the degrees of freedom (empty columns in matrix).
Returned solution is not converged.
And then I modified the standard notation to general notation in which i should put the matrix values y R matrix for H isotropic matrix. I don't know the values that i must entry in this matrix and also if is appropiate this matrix.
In summary i want confront two spheres appliying opposite pressures or forces on the top and the bottom respectively and I won't displacement on the x and y axis, only on z axis, for this reason i have selected prescribed dispalcement but i'm not sure if it is the best option and how entry the appropiate values in the matrix to converge in a solution.
Thanks a lot again Ivar
I don't understand this:
Failed to find a solution.
Segregated group 1
Singular matrix.
There are 356 void equations (empty rows in matrix) for the variable mod1.u.
at coordinates: (7.14788e-09,-2.00713e-08,4.01192e-08), (5.92103e-09,-1.98664e-08,3.72976e-08), (3.01841e-09,-2.10876e-08,4.0093e-08), (1.36218e-08,-1.63453e-08,3.99136e-08), (1.1648e-08,-1.70165e-08,3.69161e-08), ...
There are 356 void equations (empty rows in matrix) for the variable mod1.v.
at coordinates: (7.14788e-09,-2.00713e-08,4.01192e-08), (5.92103e-09,-1.98664e-08,3.72976e-08), (3.01841e-09,-2.10876e-08,4.0093e-08), (1.36218e-08,-1.63453e-08,3.99136e-08), (1.1648e-08,-1.70165e-08,3.69161e-08), ...
There are 449 void equations (empty rows in matrix) for the variable mod1.w.
at coordinates: (7.14788e-09,-2.00713e-08,4.01192e-08), (5.92103e-09,-1.98664e-08,3.72976e-08), (3.01841e-09,-2.10876e-08,4.0093e-08), (1.36218e-08,-1.63453e-08,3.99136e-08), (1.1648e-08,-1.70165e-08,3.69161e-08), ...
and similarly for the degrees of freedom (empty columns in matrix).
Returned solution is not converged.
And then I modified the standard notation to general notation in which i should put the matrix values y R matrix for H isotropic matrix. I don't know the values that i must entry in this matrix and also if is appropiate this matrix.
In summary i want confront two spheres appliying opposite pressures or forces on the top and the bottom respectively and I won't displacement on the x and y axis, only on z axis, for this reason i have selected prescribed dispalcement but i'm not sure if it is the best option and how entry the appropiate values in the matrix to converge in a solution.
Thanks a lot again Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Can you share your model, maybe I can help