static pushover analysis - results
static pushover analysis - results
Dear all,
I'm conducting pushover analysis for steel frame. As a part of verification proces I used different element's classes (FB, FBPH, DB, DBPH). During analysis i changed only element's class and number of elements per member. Results different significantly (the best fit is FBPH with PH legth 16,7%, or db with 5 or 10 elements per member) . I do not fully understand why results are so different, since all of the elements should include inelastic behviour. Should I change some additional features when changing element's classes?
Thank You in advance for help.
Best regards
michal
I'm conducting pushover analysis for steel frame. As a part of verification proces I used different element's classes (FB, FBPH, DB, DBPH). During analysis i changed only element's class and number of elements per member. Results different significantly (the best fit is FBPH with PH legth 16,7%, or db with 5 or 10 elements per member) . I do not fully understand why results are so different, since all of the elements should include inelastic behviour. Should I change some additional features when changing element's classes?
Thank You in advance for help.
Best regards
michal
Best regards,
Michal
Michal
Re: static pushover analysis - results
Hi michalbak.
As noted in the manual, DB inelastic elements often require more elements to provide accurate response, while FB inelastic elements often require little or no discretization to estimate inelastic response. These are distributed plasticity elements, unlike the FBPH and DBPH elements which are concentrated plasticity models.
I wonder what your basis for comparison is when you state that the "best fit" is the FBPH? Are you comparing to experimental results? Or hand calculations?
An interesting comparison would be to analyze using a single FB element. Then switch to DB elements and run models with 2, 3, 4, 10 elements per column. It might be that the DB elements of increasing discretization would converge to a solution (which may very well be close to the single-element FB model).
I'll be interested in hearing how you choose to solve your issue. Best of luck.
As noted in the manual, DB inelastic elements often require more elements to provide accurate response, while FB inelastic elements often require little or no discretization to estimate inelastic response. These are distributed plasticity elements, unlike the FBPH and DBPH elements which are concentrated plasticity models.
I wonder what your basis for comparison is when you state that the "best fit" is the FBPH? Are you comparing to experimental results? Or hand calculations?
An interesting comparison would be to analyze using a single FB element. Then switch to DB elements and run models with 2, 3, 4, 10 elements per column. It might be that the DB elements of increasing discretization would converge to a solution (which may very well be close to the single-element FB model).
I'll be interested in hearing how you choose to solve your issue. Best of luck.
Tim Huff
Re: static pushover analysis - results
Hi Tim,
thanks for Your replay.
I'm quite new at Seismo Soft, so for now I'd like to conduct a couple of analysis were I know the result (based on experimental tests available in literature). That's why i wrote, that the best fit is by using FBPH element's class. I carried pushover analysis with different number of elements per member using DB element class and the result was oddly different in comparison with FBPH (critical load coefficient was about a half of coefficient assess using FBPH element, changing number of DB elements did not improve the results). At this time I expected some popular with software, but after all I build the model one more time from scratch and it worked, so I suppose it was my mistake. Right now, with new model, FB elements class gives the best results. Using DB elements (with one element per member) makes the structure to stiff and as You stated, increase of discretization converge to FB results.
I hope, during further work there will no problems like that:)
Best regards,
michal
thanks for Your replay.
I'm quite new at Seismo Soft, so for now I'd like to conduct a couple of analysis were I know the result (based on experimental tests available in literature). That's why i wrote, that the best fit is by using FBPH element's class. I carried pushover analysis with different number of elements per member using DB element class and the result was oddly different in comparison with FBPH (critical load coefficient was about a half of coefficient assess using FBPH element, changing number of DB elements did not improve the results). At this time I expected some popular with software, but after all I build the model one more time from scratch and it worked, so I suppose it was my mistake. Right now, with new model, FB elements class gives the best results. Using DB elements (with one element per member) makes the structure to stiff and as You stated, increase of discretization converge to FB results.
I hope, during further work there will no problems like that:)
Best regards,
michal
Best regards,
Michal
Michal
Re: static pushover analysis - Ultimate displacement
Hello,
I am a research student at University of Geneva working on the interaction of infilled walls with RC frame elements. I am using the Sesimostruct v-7 for this purpose.
I have performed the pushover analysis to define different damage states. However I am finding difficulty in defining the ultimate displacement capacity. The pushover curve seems to go until the target displacement defined (that I have defined as 40") in the pre-processor although the collapse state is much lower (around 6.2")
Shouldn't the curve move down at collapse point instead moving straight?
I am a research student at University of Geneva working on the interaction of infilled walls with RC frame elements. I am using the Sesimostruct v-7 for this purpose.
I have performed the pushover analysis to define different damage states. However I am finding difficulty in defining the ultimate displacement capacity. The pushover curve seems to go until the target displacement defined (that I have defined as 40") in the pre-processor although the collapse state is much lower (around 6.2")
Shouldn't the curve move down at collapse point instead moving straight?
Re: static pushover analysis - results
Hi Rajen. Is the value of 6.2" for collapse displacement from a physical test? If so, what was the failure mode? Have you set performance criteria in Seismostruct to hold load or drop load once your various criteria are reached? Do failure modes from physical testing agree with the onset of damage in the model (crushing of the infill strut, column shear, etc.). Just some questions to investigate in pinpointing the discrepancy.
Best of luck.
Best of luck.
Tim Huff
Re: static pushover analysis - results
Dear Tim ,
Thank you for your response. Collapse displacement is not from physical test. Its from post processor data. In the Pre-Processor I have defined " Calculate Target Displacement" and Checked " Limit States" according to Euro-Code. After analysis "Target Displacement" dialog box in the Post-Processor gives value of displacement value at Near-collapse and other limit state.
Yes I have set performance criteria and limit values as typical value prescribed in SeismoStruct Help System.
By "Hold" or "Drop" load do you mean Strength Degradation and Type of notification? For crush core concrete I have chosen " No Residual Strength" and notification as "Stop".
Suggestions from your side will be highly appreciated. I am eager to hearing back from you.
Thank you for your response. Collapse displacement is not from physical test. Its from post processor data. In the Pre-Processor I have defined " Calculate Target Displacement" and Checked " Limit States" according to Euro-Code. After analysis "Target Displacement" dialog box in the Post-Processor gives value of displacement value at Near-collapse and other limit state.
Yes I have set performance criteria and limit values as typical value prescribed in SeismoStruct Help System.
By "Hold" or "Drop" load do you mean Strength Degradation and Type of notification? For crush core concrete I have chosen " No Residual Strength" and notification as "Stop".
Suggestions from your side will be highly appreciated. I am eager to hearing back from you.
- seismosoft
- Posts: 1192
- Joined: 06 Jul 2007, 04:55
Re: static pushover analysis - results
Hi Rajen,
The calculated target displacement is not the "Collapse displacement". It is the displacement that is roughly equal to the seismic deformation expected for the design seismic action. What one should do is to check if the members have failed or not at the calculated target displacement. For further information, please refer to the theory of pushover analysis in a good textbook. Some parts of the manual of SeismoStruct might also be very useful.
Seismosoft Support
The calculated target displacement is not the "Collapse displacement". It is the displacement that is roughly equal to the seismic deformation expected for the design seismic action. What one should do is to check if the members have failed or not at the calculated target displacement. For further information, please refer to the theory of pushover analysis in a good textbook. Some parts of the manual of SeismoStruct might also be very useful.
Seismosoft Support
Re: static pushover analysis - results
Hello,
Thank you for your response. In the code based target displacement three damage states are mentioned Damage limitation, Significant damage and Near collapse ( The structure is heavily damaged with low residual lateral strength and stiffness. Most non-structural components have collapsed. Large preferment drifts are present). This is what defined in the SeismoStruct for "Near collapse" which previously I mentioned as "Collapse displacement" (Sorry for the unaware mistake).
What I want to know is exactly how can I graphically display the collapse state or ultimate displacement in Seismostruct? As in difference other FE software,(ETABS, SAP...) the pushover curve moves downward perhaps indicating strength degradation. And if I can identify the failure mode of collapse in literature how can I incorporate in SeismoStructuct to get the output?
I would be happy to have a fruitful suggestion from your end. Looking forward to hearing back from you.
Thank you for your response. In the code based target displacement three damage states are mentioned Damage limitation, Significant damage and Near collapse ( The structure is heavily damaged with low residual lateral strength and stiffness. Most non-structural components have collapsed. Large preferment drifts are present). This is what defined in the SeismoStruct for "Near collapse" which previously I mentioned as "Collapse displacement" (Sorry for the unaware mistake).
What I want to know is exactly how can I graphically display the collapse state or ultimate displacement in Seismostruct? As in difference other FE software,(ETABS, SAP...) the pushover curve moves downward perhaps indicating strength degradation. And if I can identify the failure mode of collapse in literature how can I incorporate in SeismoStructuct to get the output?
I would be happy to have a fruitful suggestion from your end. Looking forward to hearing back from you.
- seismosoft
- Posts: 1192
- Joined: 06 Jul 2007, 04:55
Re: static pushover analysis - results
In order to effectively get a descending branch, you need to define appropriately the strength degradation in performance criteria or code-based check, e.g. 20% residual strength after shear failure.
Seismosoft Support
Seismosoft Support
Re: static pushover analysis - results
Hello,
Thanks a lot for your reply. However I have few more queries regarding the same issue. I am sorry if my queries seems stupid to you.
1) For which criterion type should I define strength degradation and shall I check type of notification as "stop" for the same?
2) And how can I identify the shear failure before the analysis as performance criteria has to be set in pre-processor.
Could you also please tell me whether I have to define a third plastic hinge at the middle height of the column to capture the sliding shear failure ( As per PhD dissertation of Crisafulli, plastic hinge develop at the middle height of the column during sliding shear failure) if so how can I do so ?
Thanks you for your information so far and I look forward to hearing back from you.
Thanks a lot for your reply. However I have few more queries regarding the same issue. I am sorry if my queries seems stupid to you.
1) For which criterion type should I define strength degradation and shall I check type of notification as "stop" for the same?
2) And how can I identify the shear failure before the analysis as performance criteria has to be set in pre-processor.
Could you also please tell me whether I have to define a third plastic hinge at the middle height of the column to capture the sliding shear failure ( As per PhD dissertation of Crisafulli, plastic hinge develop at the middle height of the column during sliding shear failure) if so how can I do so ?
Thanks you for your information so far and I look forward to hearing back from you.