material models
material models
I modeled a frame and used dynamic time-history analysis to obtain the response of that structure. First, I used stl_mp and con_ma to model reinforced concrete members. Load that I applied didn't cause members to yield - deformations were in elastic range. After that, I used el_mat to model reinforced concrete members. When I compared relative story displacements I noticed that they are different. I attached a picture (sent to <support@seismosoft.com>) , so you can see the difference and maybe explain why it occurs
-
Stelios_Antoniou
- Posts: 89
- Joined: 17 Jul 2011, 20:08
Re: material models
Hi Sandra,
The stl_mp steel model, but more importantly the con_ma concrete model are not linear even in the low deformation range. Probably this is the reason for your differences
The stl_mp steel model, but more importantly the con_ma concrete model are not linear even in the low deformation range. Probably this is the reason for your differences
Re: material models
Thank you for your answer. I thought that in con_ma model stress-strain relationship is linear for strains <0.002
-
seismosoft
- Posts: 1316
- Joined: 06 Jul 2007, 04:55
Re: material models
Yes, concrete typically features a very low tensile strength, and hence does crack at very low deformation levels.
Seismosoft Support
Seismosoft Support
Re: material models
But, When I define tensile strength in con_ma model and then run time history analysis - I get this message: 'Unable to apply the next load step. Analysis terminated'.
Re: material models
Sandra,
As others noted before, there was actually no problem with the 'con_ma' model, but rather a misunderstanding of yours on its workings, hence writing "another problem with con_ma" is factually incorrect and unfair.
The numerical convergence difficulties that you now describe may be caused by a myriad of reasons (carry out a search in this Forum on "unable to apply" and you will see what I mean), hence, again, I would be surprised if the very stable 'con_ma' material model is the one to blame.
Rui
As others noted before, there was actually no problem with the 'con_ma' model, but rather a misunderstanding of yours on its workings, hence writing "another problem with con_ma" is factually incorrect and unfair.
The numerical convergence difficulties that you now describe may be caused by a myriad of reasons (carry out a search in this Forum on "unable to apply" and you will see what I mean), hence, again, I would be surprised if the very stable 'con_ma' material model is the one to blame.
Rui
Re: material models
Sorry,not 'another'. The only problem.
In the User Manual there is a note that says: On occasions, depending on the structural model and applied loading, crack opening may introduce numerical instabilities in the analysis. By ignoring tensile resistance stability of the analysis will be achieved in easier fashion.
That is what happens in this case-when I ignore tensile resistance, I don't have problem 'unable to apply'
In the User Manual there is a note that says: On occasions, depending on the structural model and applied loading, crack opening may introduce numerical instabilities in the analysis. By ignoring tensile resistance stability of the analysis will be achieved in easier fashion.
That is what happens in this case-when I ignore tensile resistance, I don't have problem 'unable to apply'
-
seismosoft
- Posts: 1316
- Joined: 06 Jul 2007, 04:55
Re: material models
Dear Sandra,
Such instabilities are fairly common, due to the sudden crack of concrete under tension and the subsequent redistribution of forces and stresses. Some suggestions that might work in your case:
1) Read the help system what the divergence message that you get means and act accordingly, e.g. in the case of pred_ite increase the number of iterations
2) Make your convergence criteria less stringent, e.g. increase the values of tolerance
3) Increase rhe number of steps of your analysis, i.e. decrease the loading increments
4) See at what tensile strength you get divergence. It will give you a good insight of the size of the problem
SeismoSoft Support Team
Such instabilities are fairly common, due to the sudden crack of concrete under tension and the subsequent redistribution of forces and stresses. Some suggestions that might work in your case:
1) Read the help system what the divergence message that you get means and act accordingly, e.g. in the case of pred_ite increase the number of iterations
2) Make your convergence criteria less stringent, e.g. increase the values of tolerance
3) Increase rhe number of steps of your analysis, i.e. decrease the loading increments
4) See at what tensile strength you get divergence. It will give you a good insight of the size of the problem
SeismoSoft Support Team
Re: material models
1)Divergence message is fbd_ite
2)I modified: Element Loop Covergence Tolerance (1e-4, default value was 1e-7!) and Element Loop maximum iterations (300, default value was 30)
3)Problem solved
4) Thank you, suggestions were very helpful
Sandra
2)I modified: Element Loop Covergence Tolerance (1e-4, default value was 1e-7!) and Element Loop maximum iterations (300, default value was 30)
3)Problem solved
4) Thank you, suggestions were very helpful
Sandra
-
seismosoft
- Posts: 1316
- Joined: 06 Jul 2007, 04:55
Re: material models
We are glad that we helped. Just one note: your default values are not the program default values. Probably you saved a new set of settings as default. In order to restore the program defaults, click the relevant button in the Program Settings dialog box.
SeismoSoft Support Team
SeismoSoft Support Team