This is actual proper that karamba have some rules for defining the relations between components, from points to lines, lines to surfaces, surfaces to forms and volumes and so on. except from manual which you've released before, is there any references you know, we could adapt our models to, for writing algorithms more accurate?
Dear Kaz, you can run the examples them with the trial-version by decreasing the number of elements or use the free version when only beams are involved. Alternatively I can send you a one-month trial license for the full version.
I couldn't find the exact solver type you are using for the global equation system, but I guess it's a direct solver since the computation time increases heavily when exceeding 30000 nodes on my PC. Do you plan to integrate an iterative solver like PCG or similar in the next releases that steps into place when the model gets larger? For my uses this would be a great asset on your already great program because in complex models I cannot reach a mesh-independent solution without killing my PC for many hours...
the solver currently used in Karamba is a direct solver.
Did you check the memory usage when computing your model? If the machine runs out of physical memory it starts to use the hard disk which makes the solution process very slow.
A PCG solver would be a good idea. Maybe I find a good one for the next release.
I am trying to create a parametric 3D framed structure, although am struggling with the error "singluar stiffness matrix: cannot solve static problem" at the analyze model stage. Any help would be greatly appreciated. Here is the GH file grasshopper%2011.gh.
Hi Felicity...this would be better on the discussion forum than here. But it was because your vertical members weren't broken down, they were all one element. You needed to break them down to intersect with the horizontals. See attached for a method of doing this (disclaimer - there are probably slicker ways of doing the intersect). I believe the model runs now. grasshopper11_jjs.gh
This is why you need to understand the principles of structural engineering. You have an unstable structure. It's dangerous to just assume Karamba or any structural analysis algorithm is giving you correct answers. I suggest you look first at your support conditions, then your material, then your section sizes then your applied load. You should then also so some sanity checks by Doing some hand calculations using first principles of static structures
Dear Huax, did you try to evaluate the eigen-modes of your structure? The first eigen-mode shows how the structure moves as a rigid body in case it is kinematic.
Best, Clemens
P.S.: Would you mind opening a new discussion on the subject? It would be better readable for others.
Hi Patricio, if you flatten the 'Elem'-input at the 'Assemble'-component the definition works. The triangular shell elements have linear displacement interpolations whereas the beam deflections are exact. In order to get correct results you should refine the shell mesh.
Hello, succeeds in creating the mesh to the slab, and built the beam segment, but when I see the deformations are not expected because the beam is deformed as the slab.
Thanks for the help
PS: maybe I'm using the program for a type of structure that is not the most appropriate, as I saw in the examples of other structures. But this type of structure is that students taught
Perfect, that is just what I needed, thank you very much for your help, I'll keep working and any questions write again, it is very useful for students to understand the structural performance.
I was doing a rough estimate of the deformation, and I can not achieve the same result with Karamba. When I make a rough estimate of the result with Karamba beams and mine are very similar, I think the problem is when I connect the shell, because there are no similar results.
I sent the GH file, and an image of the calculation
The structure is concrete The result I get is 0.58cm
try to increase the number of shell elements. As mentioned in the manual they are linear elements. A mesh that is too coarse leads to a response which is stiffer than the real structure.
Best,
Clemens
P.S.: Could you please open a new discussion on the topic? These discussions are easier to find and follow for other people who face a similar problem.
I need your help! I have a very simple curved geometry, like an arch, composed by several beams. This is a not funicular geometry and there are huge bending moments. I do not understand:
1) if the Cross Section Optimization component takes into account buckling for members in bending and axial compression (6.3.3 - Eurocode 3).
2) if the buckling length corresponds to each beam or to the length of the whole arch or to the distance between supports (local or global buckling.
If instead of an arch, there is a more complex structure, i.e. an arch bridge with tendons and deck, what is the buckling length of the arch adopted by the Cross Section Optimization component?
the Cross Section Optimization component takes into account buckling for members (see the appendix of the Karamba manual). The assumed buckling length is the distance between two nodes that connect to more than three beams or to a support or double that distance in case one end is free. The buckling lengths can be set manually via the 'ModifyBeam'-component. If you have 'huge' bending moments in your structure then buckling will have little effect on the result.
Sometimes so called global buckling can govern a design (think e.g. of a compressed girder of a truss structure). In such cases one has to guess the buckling length or do a buckling analysis (which will be possible in Karamba 1.1.0).
in Karamba 1.0.5 the cross section optimization algorithm only considers normal forces for buckling. The next version (Karamba 1.1.0) takes full account of the EC3 formulas including lateral torsional buckling and interaction values according to annex B.
Is there any way to change the eccentricity of the post-tension load? I changed the ecce-loc input but that only changes the centroid of where the point load acts -- the post-tension load seems to be unaffected.
2. How do I get a filled trapezoid cross section? It appears to be hollow at the moment.
Thanks for your response. However, I'm still having trouble with the ecce-loc. I want to change ONLY the eccentricity of the prestress load , and not offset the other uniform loads. Is there a way to do that?
The problem with this method is that the model won't consider internal stresses. The moving has to take place after the point is converted to a support variable.
Hi, Rafael! I did, but i didn't manage to get what i wanted. I used the prescribed displacement component with an input vector and a force acting on the points of the prescribed displacement hoping that once the amplitude of the vector is reached, the points don't move any further, but they di surpass the length of the vector...
New tutorials showing workflow between Karamba and Revit using Geometry Gym and Dynamo are up on our youtube and Grasshopper pages! The full definitions will follow shortly!
Karamba 1.2.0 is now available! The latest release contains a new solver that is twice as fast and uses less memory. New components such as the BESO for Shells and Deformation Energy have been introduced. The entire list of updates and bug-fixes can be found at http://www.grasshopper3d.com/group/karamba/page/new-features-and-bu.... Look out for some new examples coming soon.
the image is similar to the setup in the example on our website. You can adjust the element felting settings to and add additional parameters to achieve different results.
Hello Alexia, Karamba is a Finite Element toolkit and thus does not depend on national specifications. The only exeption is the cross section optimization algorithm which is based on Eurocode 3.
If I use the last version of Karamba and I need to know the best sections of the beams do I need to use galapagos or is it ok to use the option of the best beam?
Karamba3D
"karamba.ini" is located in the folder "Karamba" which can be found in the installation directory of Karamba.
Jan 12, 2014
Mahdi Soheyli Fard
Hi there,
long time no comments!
Dear Clemens,
This is actual proper that karamba have some rules for defining the relations between components, from points to lines, lines to surfaces, surfaces to forms and volumes and so on. except from manual which you've released before, is there any references you know, we could adapt our models to, for writing algorithms more accurate?
Thanks,
Mahdi
Mar 19, 2014
Kaz
All examples files are restricted in trial version , how to test it ?
Mar 25, 2014
Karamba3D
Dear Kaz, you can run the examples them with the trial-version by decreasing the number of elements or use the free version when only beams are involved. Alternatively I can send you a one-month trial license for the full version.
Best, Clemens
Mar 25, 2014
Kaz
it would be better to test full version , thank you for your cooperation , here is my email : blumen.sy@gmail.com
thanks again
Mar 27, 2014
pablo
Dear Clemens,
Could I enjoy that one-month trial license? I need to try beam joints components.
Thank you very much.
May 21, 2014
Robert Naguschewski
Hi karamba-team,
I couldn't find the exact solver type you are using for the global equation system, but I guess it's a direct solver since the computation time increases heavily when exceeding 30000 nodes on my PC. Do you plan to integrate an iterative solver like PCG or similar in the next releases that steps into place when the model gets larger? For my uses this would be a great asset on your already great program because in complex models I cannot reach a mesh-independent solution without killing my PC for many hours...
Thank you,
Robert
Aug 5, 2014
Karamba3D
Hi Robert,
the solver currently used in Karamba is a direct solver.
Did you check the memory usage when computing your model? If the machine runs out of physical memory it starts to use the hard disk which makes the solution process very slow.
A PCG solver would be a good idea. Maybe I find a good one for the next release.
Best,
Clemens
Aug 5, 2014
Felicity Stewart
Hi All,
I am trying to create a parametric 3D framed structure, although am struggling with the error "singluar stiffness matrix: cannot solve static problem" at the analyze model stage. Any help would be greatly appreciated. Here is the GH file grasshopper%2011.gh.
Aug 12, 2014
jjsolly
Hi Felicity...this would be better on the discussion forum than here. But it was because your vertical members weren't broken down, they were all one element. You needed to break them down to intersect with the horizontals. See attached for a method of doing this (disclaimer - there are probably slicker ways of doing the intersect). I believe the model runs now. grasshopper11_jjs.gh
Aug 13, 2014
Felicity Stewart
Thanks! I will make sure to direct any further queries to the discussion forum, but my many thanks for getting it working :)
Aug 13, 2014
Steve Lewis
Regarding your comment below.
This is why you need to understand the principles of structural engineering. You have an unstable structure. It's dangerous to just assume Karamba or any structural analysis algorithm is giving you correct answers. I suggest you look first at your support conditions, then your material, then your section sizes then your applied load. You should then also so some sanity checks by Doing some hand calculations using first principles of static structures
Sep 27, 2014
Karamba3D
Dear Huax,
I tried out your file but could not reproduce the error. Du you use the latest version of Karamba?
Best, Clemens
Sep 28, 2014
Karamba3D
Dear Huax, did you try to evaluate the eigen-modes of your structure? The first eigen-mode shows how the structure moves as a rigid body in case it is kinematic.
Best, Clemens
P.S.: Would you mind opening a new discussion on the subject? It would be better readable for others.
Sep 29, 2014
Llordella Patricio
8-10-14losa.gh
hello, I'm working on this model. I would like to know how to connect the slab with beams, so that the whole work as one element.
Thank you
Oct 7, 2014
Karamba3D
Hello Patricio, divide the beams in such a way that each boundary vertex of the shell becomes an endpoint of a beam segment.
Best, Clemens
Oct 7, 2014
Llordella Patricio
Hi Clemens,
I did what you suggested but now assemble element doesn´t work properly. Could you please tell me how to fix it? Thanks in advance, Patricio
8-10-14losa%20cadena.gh
Oct 8, 2014
Karamba3D
Hi Patricio,
if you flatten the 'Elem'-input at the 'Assemble'-component the definition works. The triangular shell elements have linear displacement interpolations whereas the beam deflections are exact. In order to get correct results you should refine the shell mesh.
Best, Clemens
Oct 8, 2014
Llordella Patricio
Hello, succeeds in creating the mesh to the slab, and built the beam segment, but when I see the deformations are not expected because the beam is deformed as the slab.
Thanks for the help
PS: maybe I'm using the program for a type of structure that is not the most appropriate, as I saw in the examples of other structures. But this type of structure is that students taught
best regards
Patricio
9-10-14%20Example%201.gh
Oct 9, 2014
Karamba3D
You could use the 'Mesh Edges'-component to retrieve the naked edges and turn them into beams - see attached file: 91014Example1_cp.gh
Best regards,
Clemens
Oct 9, 2014
Llordella Patricio
Perfect, that is just what I needed, thank you very much for your help, I'll keep working and any questions write again, it is very useful for students to understand the structural performance.
best regards
Patricio
Oct 10, 2014
Llordella Patricio
Dear clemens
I was doing a rough estimate of the deformation, and I can not achieve the same result with Karamba.
When I make a rough estimate of the result with Karamba beams and mine are very similar, I think the problem is when I connect the shell, because there are no similar results.
I sent the GH file, and an image of the calculation
The structure is concrete
The result I get is 0.58cm
thank you
Patricio
15-10-14%20Example.gh
Oct 15, 2014
Karamba3D
Dear Patricio,
try to increase the number of shell elements. As mentioned in the manual they are linear elements. A mesh that is too coarse leads to a response which is stiffer than the real structure.
Best,
Clemens
P.S.: Could you please open a new discussion on the topic? These discussions are easier to find and follow for other people who face a similar problem.
Oct 21, 2014
Llordella Patricio
Dear clemens
Thanks for the help, I just created a topic discussion and add any comments that we have done so far.
Best regards
Patricio
Oct 22, 2014
Leonardo
Hi Clemens,
I need your help! I have a very simple curved geometry, like an arch, composed by several beams. This is a not funicular geometry and there are huge bending moments. I do not understand:
1) if the Cross Section Optimization component takes into account buckling for members in bending and axial compression (6.3.3 - Eurocode 3).
2) if the buckling length corresponds to each beam or to the length of the whole arch or to the distance between supports (local or global buckling.
If instead of an arch, there is a more complex structure, i.e. an arch bridge with tendons and deck, what is the buckling length of the arch adopted by the Cross Section Optimization component?
Many thanks,
Leonardo
Jan 7, 2015
Karamba3D
Hi Leonardo,
the Cross Section Optimization component takes into account buckling for members (see the appendix of the Karamba manual). The assumed buckling length is the distance between two nodes that connect to more than three beams or to a support or double that distance in case one end is free. The buckling lengths can be set manually via the 'ModifyBeam'-component.
If you have 'huge' bending moments in your structure then buckling will have little effect on the result.
Sometimes so called global buckling can govern a design (think e.g. of a compressed girder of a truss structure). In such cases one has to guess the buckling length or do a buckling analysis (which will be possible in Karamba 1.1.0).
Best,
Clemens
Jan 7, 2015
Leonardo
Clemens, thanks for your quick answer.
I have carefully read the appendix and the way for calculating the buckling length is clear. However I have still one questions:
seems to be that adopting Eurocode3 (eq. 6.47) only the normal forces are compared with the critical buckling load. Am I right?
Can we want take into account both, bending and axial compression, as suggested by Eurocode3 (Eq. 6.3.3)?
Many thanks,
Leonardo
Jan 8, 2015
Karamba3D
Leonardo:
in Karamba 1.0.5 the cross section optimization algorithm only considers normal forces for buckling. The next version (Karamba 1.1.0) takes full account of the EC3 formulas including lateral torsional buckling and interaction values according to annex B.
Best,
Clemens
Jan 8, 2015
Nicholas Rawitscher
Hey guys! do you know where to find the "point-load" component in the latest version of Karamba? thank you for your help
Mar 31, 2015
Elisa Cheung
Hey guys, I'm having two issues.
Is there any way to change the eccentricity of the post-tension load? I changed the ecce-loc input but that only changes the centroid of where the point load acts -- the post-tension load seems to be unaffected.
2. How do I get a filled trapezoid cross section? It appears to be hollow at the moment.
Thanks for your help in advance!
Apr 13, 2015
Karamba3D
Hi Elisa,
in Karamba 1.1.0 the post-tension load is attached to the eccentric beam element (see PreTensionLoad_EccentricBeam.gh).
The trapezoid cross section is filled (when rendered it looks hollow though). The box cross section is hollow.
Best,
Clemens
Apr 13, 2015
Elisa Cheung
Clemens,
Thanks for your response. However, I'm still having trouble with the ecce-loc. I want to change ONLY the eccentricity of the prestress load , and not offset the other uniform loads. Is there a way to do that?
Apr 14, 2015
Karamba3D
Elisa, you could try to place two elements in the same spot. One with, one without eccentricity and add corresponding loads to them.
Best, C.
Apr 15, 2015
Andrei Raducanu
Can curved folding simulations be done using Karamba?

Something along these lines :
Apr 18, 2015
Andrei Raducanu
What I don't get is why in the simply supported plate and beam examples one of the ends doesn't move horizontally when the beam/plate gets bent.
It does have the freedom to do so and in order to keep the length constant it should do so, yet it doesn't. Please answer if You can
Apr 18, 2015
Elisa Cheung
Clemens,
What do you mean by placing two elements in the same spot? Wouldn't that give me two separate beams?
Apr 19, 2015
Andrei Raducanu
Hi all,
Is there a way to move supports by a certain distance in a certain direction?
(say i have a shell and i want to move one of its vertexes and see what stresses appear)
Apr 19, 2015
Rafael Pastrana
Hi Andrei,I guess you would need to move the point that defines the aforementioned support in order to do that.
Best,
Rafael
Apr 20, 2015
Andrei Raducanu
Hi, Rafael,
The problem with this method is that the model won't consider internal stresses. The moving has to take place after the point is converted to a support variable.
Apr 20, 2015
Rafael Pastrana
Hi Andrei, have you had a look to the "prescribed displacement" component? I guess this one will help you out with what you want. cheers, Rafael.
Apr 20, 2015
Andrei Raducanu
Hi, Rafael! I did, but i didn't manage to get what i wanted. I used the prescribed displacement component with an input vector and a force acting on the points of the prescribed displacement hoping that once the amplitude of the vector is reached, the points don't move any further, but they di surpass the length of the vector...
Apr 20, 2015
Karamba3D
Dear all,
in case of questions please open new discussions (see above). They are easier to find and to follow than entries on the comment wall.
Best,
Clemens
Sep 19, 2015
Karamba3D
New tutorials showing workflow between Karamba and Revit using Geometry Gym and Dynamo are up on our youtube and Grasshopper pages! The full definitions will follow shortly!
Jan 8, 2016
Karamba3D
Hi all!
Karamba 1.2.0 is now available! The latest release contains a new solver that is twice as fast and uses less memory. New components such as the BESO for Shells and Deformation Energy have been introduced. The entire list of updates and bug-fixes can be found at http://www.grasshopper3d.com/group/karamba/page/new-features-and-bu.... Look out for some new examples coming soon.
It can be downloaded from our website or from food4rhino. Enjoy!
Your Karamba Team
Jul 8, 2016
Marco Pellegrino
Hi All!
does anyone know how to make this kind of image?
I saw different image like this made from Robert ( Hi Robert ) but I wasn't able.
Best,
Marco
Sep 12, 2016
Karamba3D
Hi Marco,
the image is similar to the setup in the example on our website. You can adjust the element felting settings to and add additional parameters to achieve different results.
Matt
Sep 13, 2016
Alexia Ch
second order theory (Th.II) could you pls help me understand what this theory is called in Italian? is Karamba ok to do the analysis in Italy?
Oct 31, 2016
Karamba3D
Hello Alexia, Karamba is a Finite Element toolkit and thus does not depend on national specifications. The only exeption is the cross section optimization algorithm which is based on Eurocode 3.
Nov 1, 2016
Alexia Ch
If I use the last version of Karamba and I need to know the best sections of the beams do I need to use galapagos or is it ok to use the option of the best beam?
Nov 6, 2016
Balaji Visvanath
Hello Team Karamba,
I am new to Karamba and have just purchased a PRO license (waiting for activation).
I am currently working on Voronoi Gridshell structures for my Master Thesis and came across a very relevant example on your side. The example is http://www.karamba3d.com/examples/hard/projected-voronoi-pattern-gr....
I am very interested in knowing more about it. Is there any related tutorial ? (i could not find any on the site)
Kind regards,
Balaji
Nov 9, 2016