material
Moderators: silvia, selimgunay, Moderators
Re: material
Unfortunately Opensees doesn't consider any hysteresis model. here is the link that proves my saying:
http://opensees.berkeley.edu/community/ ... del#p98992
Although it has been viewed for more than 200 times, nobody (even software developers!) has given any answer to this question. Anyway, I highly recommend you to use IDARC instead of Opensees. It considers a variety of hysteresis models such as clough, and is much more powerful than opensees. However, it restricts you to RC frames only.
http://opensees.berkeley.edu/community/ ... del#p98992
Although it has been viewed for more than 200 times, nobody (even software developers!) has given any answer to this question. Anyway, I highly recommend you to use IDARC instead of Opensees. It considers a variety of hysteresis models such as clough, and is much more powerful than opensees. However, it restricts you to RC frames only.
Re: material
The topic did not get a response because it makes no sense! How is it that OpenSees has been used extensively for almost 15 years for simulating cyclic earthquake response of structural and geotechnical systems (not just RC frames) all the while not having capabilities to model hysteretic behavior???
It is true however that OpenSees is not well documented, perhaps therein lies the problem you are having. There is HysteresticMaterial and several Clough, pinching, bilinear models implemented under the uniaxialMaterial framework. If you are looking for a specific model, post it here.
It is true however that OpenSees is not well documented, perhaps therein lies the problem you are having. There is HysteresticMaterial and several Clough, pinching, bilinear models implemented under the uniaxialMaterial framework. If you are looking for a specific model, post it here.
Re: material
Dear Scott,
I cannot understand what you mean by "it makes no sense ..."!
Imagine we have a 2D frame and we want to perform a nonlinear dynamic analysis on it. I want the behavior of THE WHOLE STRUCTURE to be like Clough's model, while you might want it to behave as Takeda's model. Definitely, there are lots of programs that perform nonlinear analysis for us. Among them, I can mention "Abaqus", "SAP2000", and "IDARC". And all of them permit the user to utilize a variety of hysteresis models. For instance:
1) If you want to use "Abaqus" these hysteresis models are available:
Yeoh model, Ogden model, ...
2) If you want to use "SAP2000" these hysteresis models are available:
Pivot model, Takeda model, ...
3) And finally, if you want to use "IDARC" there are lots of hysteresis models available to user:
Park model, Takeda model, clough model, Maxwell model, ...
But what if we apply "OpenSees" instead? Does it have the capability of using different hysteresis models? If it has, just let us know.
And let me add something to this. This is obvious that we can define all the abovementioned hysteresis models in "OpenSees" by using HysteresticMaterial, or maybe by using uniaxialMaterial, but the question is that HOW do you assign a certain hysteretic behavior to the whole structural model.
Warm regards
I cannot understand what you mean by "it makes no sense ..."!
Imagine we have a 2D frame and we want to perform a nonlinear dynamic analysis on it. I want the behavior of THE WHOLE STRUCTURE to be like Clough's model, while you might want it to behave as Takeda's model. Definitely, there are lots of programs that perform nonlinear analysis for us. Among them, I can mention "Abaqus", "SAP2000", and "IDARC". And all of them permit the user to utilize a variety of hysteresis models. For instance:
1) If you want to use "Abaqus" these hysteresis models are available:
Yeoh model, Ogden model, ...
2) If you want to use "SAP2000" these hysteresis models are available:
Pivot model, Takeda model, ...
3) And finally, if you want to use "IDARC" there are lots of hysteresis models available to user:
Park model, Takeda model, clough model, Maxwell model, ...
But what if we apply "OpenSees" instead? Does it have the capability of using different hysteresis models? If it has, just let us know.
And let me add something to this. This is obvious that we can define all the abovementioned hysteresis models in "OpenSees" by using HysteresticMaterial, or maybe by using uniaxialMaterial, but the question is that HOW do you assign a certain hysteretic behavior to the whole structural model.
Warm regards
Re: material
You can use any uniaxialMaterial in a zero length element in order to model an SDOF or MDOF system. The uniaxialMaterials would represent the story force-deformation relationship. I assume this is what you mean by the "whole structural model".
Take a look in SRC/material/uniaxial/ and SRC/material/uniaxial/drain/ to see what models are available.
Take a look in SRC/material/uniaxial/ and SRC/material/uniaxial/drain/ to see what models are available.
Re: material
Dear Scott,
First of all, I should say I believe in your keen knowledge in opensees and also in structural engineering. The reason that I continue this discussion is only to debate on whether opensees has such capabilities or not. And I really appreciate your effort.
Anyway, what I mean by the "whole structural model" is a "real model". For instance, a RC frame that you model by using ordinary commands in opensees, and by following such steps:
1- Define materials:
uniaxialMaterial Concrete01 ...
uniaxialMaterial Steel02 ...
uniaxialMaterial ConfinedConcrete01 ...
2-Define sections:
section Fiber ... patch quad ... layer straight ...
3-Define Elements:
element dispBeamColumn ...
When the model is built and ready to be analyzed, before performing nonlinear analysis, the programs I mentioned in my last post (Abaqus, IDARC, ...) give you the opportunity to decide on how you want your model to behave under cycling loadings. Therefore, you can assign a certain hysteresis model to your frame building, and you will expect your model to behave like that hysteresis model under cyclic loadings.
Apparently, this is completely different from using just ONE uniaxialMaterial to represent the shear force vs. roof displacement relationship, or to use zero length elements in order to model an SDOF or MDOF system. When you go in this way, you are not modeling a real model, and you are not assigning any hysteresis model (such as clough, Takeda, ...) to a real model! Although the latest way is suitable for simulations, this is distinctive to the procedure I explained above. I just want to know, after you build a real model by using different materials and different sections, is it possible to assign a certain hysteresis behavior to the model in opensees, and if so, how?!
Kind regards
First of all, I should say I believe in your keen knowledge in opensees and also in structural engineering. The reason that I continue this discussion is only to debate on whether opensees has such capabilities or not. And I really appreciate your effort.
Anyway, what I mean by the "whole structural model" is a "real model". For instance, a RC frame that you model by using ordinary commands in opensees, and by following such steps:
1- Define materials:
uniaxialMaterial Concrete01 ...
uniaxialMaterial Steel02 ...
uniaxialMaterial ConfinedConcrete01 ...
2-Define sections:
section Fiber ... patch quad ... layer straight ...
3-Define Elements:
element dispBeamColumn ...
When the model is built and ready to be analyzed, before performing nonlinear analysis, the programs I mentioned in my last post (Abaqus, IDARC, ...) give you the opportunity to decide on how you want your model to behave under cycling loadings. Therefore, you can assign a certain hysteresis model to your frame building, and you will expect your model to behave like that hysteresis model under cyclic loadings.
Apparently, this is completely different from using just ONE uniaxialMaterial to represent the shear force vs. roof displacement relationship, or to use zero length elements in order to model an SDOF or MDOF system. When you go in this way, you are not modeling a real model, and you are not assigning any hysteresis model (such as clough, Takeda, ...) to a real model! Although the latest way is suitable for simulations, this is distinctive to the procedure I explained above. I just want to know, after you build a real model by using different materials and different sections, is it possible to assign a certain hysteresis behavior to the model in opensees, and if so, how?!
Kind regards
Re: material
If all you are doing is assigning hysteresis models to your building in order to give a response you want (is that really real?), then you can use zeroLength elements. Real dimensions, member sizes, etc. become irrelevant when you assign hysteretic behavior in the manner you describe.
Please do not trash OpenSees (or any other software) until you fully explore its capabilities. Those of us on the message board can help if you are looking for a specific model.
Please do not trash OpenSees (or any other software) until you fully explore its capabilities. Those of us on the message board can help if you are looking for a specific model.
Re: material
Dear Scott,
I didn't mean to trash any software. I just said what I was thinking. However, I'm thankful to you for your patience, and changing my mind!
Regards
I didn't mean to trash any software. I just said what I was thinking. However, I'm thankful to you for your patience, and changing my mind!
Regards
Re: material
I have to agree that the questions make no sense...
The beauty of Opensees (and open source codes in general) is, that gives you the option to easily create something if it's not there
. As I can understand from the context is, that in some similar software you define the lumped macro models with line elements with the hysteresis, whereas in opensees you need two zerolength elements (or one in the case of cantilever) and a line element connecting them (i.e. you define the macro element by yourself, although one could easily automate that. The neonlinear bending behaviour is "hidden" in plastic hinges and elastic in the beam/column. I hope this clarifies the confusion. By setting the parameters of uniaxial hysteretic material you can mimic either Park, Takeda or Clough (not sure about the stiffness, strength degradation).
google "nonlinear plastic hinge opensees" or check http://web.engr.oregonstate.edu/~mhscot ... E_2008.pdf and references there in
Confusing/mixing the use of hyperelastic (yeoh and ogden) material with uniaxial clough hysteretic model also makes no sense.
The beauty of Opensees (and open source codes in general) is, that gives you the option to easily create something if it's not there
. As I can understand from the context is, that in some similar software you define the lumped macro models with line elements with the hysteresis, whereas in opensees you need two zerolength elements (or one in the case of cantilever) and a line element connecting them (i.e. you define the macro element by yourself, although one could easily automate that. The neonlinear bending behaviour is "hidden" in plastic hinges and elastic in the beam/column. I hope this clarifies the confusion. By setting the parameters of uniaxial hysteretic material you can mimic either Park, Takeda or Clough (not sure about the stiffness, strength degradation). google "nonlinear plastic hinge opensees" or check http://web.engr.oregonstate.edu/~mhscot ... E_2008.pdf and references there in
Confusing/mixing the use of hyperelastic (yeoh and ogden) material with uniaxial clough hysteretic model also makes no sense.