Dear Seisomstruct support team,
My name is Hani Akkari, I ‘m preparing my master’s degree on the performance of concentrically braced steel frames under seismic loading, I prepared a sample model which is a 2 story concentrically braced steel frame. I went through the help of your software and modeled the steel braces using the constraint commend (equal degree of freedom, slaves and master nodes)and I loaded the first floor at the joint (na2) by an horizontal force of 3 KN and the joint (na3) on the second floor by a horizontal force of 10 KN. I used the pushover analysis to get the pushover curve as per the attached figure 1. My problem is the following:
In the output we have the so called load factor versus displacement or the base shear versus displacement. The definition of the load factor in the help is that if you multiply that factor at a certain step by the load pattern entered(in my case 10KN+3KN =13KN ) You will get as a result the force that caused the displacement at that step , so that force should be equal to the base shear if I’m not wrong. So my problem is at a certain step the product of the load factor by the applied load is not equal to the base shear at that step?
I will give an example from my model : if I plot load factor against the displacement at joint (na3), at the step where the displacement is 0.031m the load factor is 460, so at that step the force causing this displacement is 460x13KN=5980 KN, But when I plot the base shear against the displacement at 0.031m I have 630 KN as base shear?
Please can you explain to what is the problem?
Attached is the model and some shots load factor an base shear.
model :
https://dl.dropbox.com/u/33925535/model-Hani.spf
snapshots:
https://dl.dropbox.com/u/33925535/snaps ... 0loads.JPG
https://dl.dropbox.com/u/33925535/base% ... cement.JPG
https://dl.dropbox.com/u/33925535/load% ... cement.JPG
regards,
monitoring performance of steel braces
Re: monitoring performance of steel braces
Hello and greetings hani. I am a little confused as to whay you are proposing to use so many coincident nodes at each intersectin point.
I may be missing some feature you wish to emulate, but I believe you could do this with no coincident nodes, using a total 8 nodes. And eliminate all constraints.
I believe it is the constraints which are causing the issue. It seems that if you simplify the model as indicated, the load factor x 13 will equal the base shear.
I have pasted the modified nodes and element connectivities below. See if this will capture the response you wish to model.
/&Start of list&/
na1 0. 0. 0. structural
nb1 4. 0. 0. structural
na2 0. 0. 3.6 structural
nb2 4. 0. 3.6 structural
nsn1001 8 0. 12 non-structural
ma1 2 0. 1.8 structural
ma2 2. 0. 5.4 structural
na3 0. 0. 7.2 structural
nb3 4. 0. 7.2 structural
/&End of list&/
/&Start of list&/
c1a COLUMNS na1 na2 nb1
c1b COLUMNS nb1 nb2 na1
b1 BEAM na2 nb2 na1
b2 BEAM na3 nb3 na1
c2a COLUMNS na2 na3 nb2
c2b COLUMNS nb2 nb3 na2
t1a TRUSS(FRAME) na1 ma1 na2
t1b TRUSS(FRAME) na2 ma1 nb2
t1c TRUSS(FRAME) nb2 ma1 na2
t1d TRUSS(FRAME) nb1 ma1 na1
t2a TRUSS(FRAME) na2 ma2 nb2
t2b TRUSS(FRAME) na3 ma2 na2
t2c TRUSS(FRAME) nb3 ma2 nb2
t2d TRUSS(FRAME) nb2 ma2 na2
/&End of list&/
I may be missing some feature you wish to emulate, but I believe you could do this with no coincident nodes, using a total 8 nodes. And eliminate all constraints.
I believe it is the constraints which are causing the issue. It seems that if you simplify the model as indicated, the load factor x 13 will equal the base shear.
I have pasted the modified nodes and element connectivities below. See if this will capture the response you wish to model.
/&Start of list&/
na1 0. 0. 0. structural
nb1 4. 0. 0. structural
na2 0. 0. 3.6 structural
nb2 4. 0. 3.6 structural
nsn1001 8 0. 12 non-structural
ma1 2 0. 1.8 structural
ma2 2. 0. 5.4 structural
na3 0. 0. 7.2 structural
nb3 4. 0. 7.2 structural
/&End of list&/
/&Start of list&/
c1a COLUMNS na1 na2 nb1
c1b COLUMNS nb1 nb2 na1
b1 BEAM na2 nb2 na1
b2 BEAM na3 nb3 na1
c2a COLUMNS na2 na3 nb2
c2b COLUMNS nb2 nb3 na2
t1a TRUSS(FRAME) na1 ma1 na2
t1b TRUSS(FRAME) na2 ma1 nb2
t1c TRUSS(FRAME) nb2 ma1 na2
t1d TRUSS(FRAME) nb1 ma1 na1
t2a TRUSS(FRAME) na2 ma2 nb2
t2b TRUSS(FRAME) na3 ma2 na2
t2c TRUSS(FRAME) nb3 ma2 nb2
t2d TRUSS(FRAME) nb2 ma2 na2
/&End of list&/
Tim Huff
-
hani.akkari
- Posts: 18
- Joined: 20 Aug 2010, 22:00
Re: monitoring performance of steel braces
dear Huffte,
thanks for trying to solve my issue , the point that you missed is that the steel braces has to be relesed at their end (to have only axial force inside ) so all braces will be connected to a slave node attached to the master node but with free rotation that's why i created these additional slave nodes (from the constraint command).
i hope that these additional information will help you to re-investigate in my subject again.
Regards,
Hani
thanks for trying to solve my issue , the point that you missed is that the steel braces has to be relesed at their end (to have only axial force inside ) so all braces will be connected to a slave node attached to the master node but with free rotation that's why i created these additional slave nodes (from the constraint command).
i hope that these additional information will help you to re-investigate in my subject again.
Regards,
Hani
Re: monitoring performance of steel braces
Hani, can you consider using inelastic truss type elements for your braces?
This will automatically hinge the ends of the braces. You will receive indication of a zero-diagonal stiffness value at the brace intersection nodes, but the analysis will still run and the results do look logical.
This will automatically hinge the ends of the braces. You will receive indication of a zero-diagonal stiffness value at the brace intersection nodes, but the analysis will still run and the results do look logical.
Tim Huff
Re: monitoring performance of steel braces
In addition, you could assign an artificially high material strength to the brace material if, for some reason, you wanted to make certain no inelastic behavior in the braces is considered in the analysis.
On the other hand, the use of the inelastic truss element would identify any yielding in the braces if you use the actual material strength.
On the other hand, the use of the inelastic truss element would identify any yielding in the braces if you use the actual material strength.
Tim Huff
-
hani.akkari
- Posts: 18
- Joined: 20 Aug 2010, 22:00
Re: monitoring performance of steel braces
Dear Huffte,
actually my thesis is on monitoring the performance of steel braces, this means the formation of plastic hinges (performance criteria as per FEMA 356), i don't think that this is possible with the inelastic truss element as you can't assign to it the performance levels(as i understand , with the inelastic truss element we can monitor the global performance of the hole BRACE FRAME, but not the local issues.) i didnt understand your point of using an artificially high material strength to the brace, in what this can help me?. can you inform me please as i said before: if i want to monitor the steel braces performane, what will be the best solution to do?.
best Regards,
Hani
actually my thesis is on monitoring the performance of steel braces, this means the formation of plastic hinges (performance criteria as per FEMA 356), i don't think that this is possible with the inelastic truss element as you can't assign to it the performance levels(as i understand , with the inelastic truss element we can monitor the global performance of the hole BRACE FRAME, but not the local issues.) i didnt understand your point of using an artificially high material strength to the brace, in what this can help me?. can you inform me please as i said before: if i want to monitor the steel braces performane, what will be the best solution to do?.
best Regards,
Hani
Re: monitoring performance of steel braces
Hani, I am still just a bit confused. If you are studying the behavior of braced frames, can I assume that your reference to plastic hinges refers to axial hinges uin the braces as opposed to felxural hinges.
You will be able to obtain both axial force vs. axial displacement plots for the individual braces and global story shear vs. nodal displacement plots when inelastic truss elements are used. In fact, on page 5-44 of FEMA 356, I notice that the acceptance criteria for braces (when a nonlinear procedure is employed) is axial deformation in the braces. Are you using this criteria from FEMA 356 or some other criteria?
I may still be missing something Hani, but I do have difficulty in seeing what the modeling of frame elements with restraints and coincident nodes provides in the way of advantages over the simpler inelastic truss elements.
I hope this correspondence will eventually clear things up for you Hani. Best of luck in your interesting work.
You will be able to obtain both axial force vs. axial displacement plots for the individual braces and global story shear vs. nodal displacement plots when inelastic truss elements are used. In fact, on page 5-44 of FEMA 356, I notice that the acceptance criteria for braces (when a nonlinear procedure is employed) is axial deformation in the braces. Are you using this criteria from FEMA 356 or some other criteria?
I may still be missing something Hani, but I do have difficulty in seeing what the modeling of frame elements with restraints and coincident nodes provides in the way of advantages over the simpler inelastic truss elements.
I hope this correspondence will eventually clear things up for you Hani. Best of luck in your interesting work.
Tim Huff
-
hani.akkari
- Posts: 18
- Joined: 20 Aug 2010, 22:00
Re: monitoring performance of steel braces
Dear Huffte,
i'm using the FEMA 356 and the acceptance criteria is as you said the axial deformation, but the fact that i didn't used the inelastic truss element is that you can't set performance criteria for it as for the frame element(giving for each strain a color to notify when the strain in the truss element reach the entered value), the curve for the brace that you are talking about will not give me the exact value of the strain , maybe the value that i want to monitor will come in between 2 sucessive values. so what do you think?
regards,
i'm using the FEMA 356 and the acceptance criteria is as you said the axial deformation, but the fact that i didn't used the inelastic truss element is that you can't set performance criteria for it as for the frame element(giving for each strain a color to notify when the strain in the truss element reach the entered value), the curve for the brace that you are talking about will not give me the exact value of the strain , maybe the value that i want to monitor will come in between 2 sucessive values. so what do you think?
regards,
Re: monitoring performance of steel braces
Hani,
It seems that a strain criterion for a purely axial member should be easily converted to a deformation criterion. The load deflection behavior for an inelastic truss element, I believe, will be very simple and bi-linear.
I think you might be well-served in doing your post-processing external to SeismoStruct.
I'm still not sure what was going on with the joint constraint model you first proposed. If the post-processing you wish to do is more convenient for that model, then you could run both models and try to determine which number, load factor or base shear, is the appropriate one from the constrained model by comparison to the simpler model with inelastic trusses.
I think you are getting close to a solution though.
Kind regards,
huffte
It seems that a strain criterion for a purely axial member should be easily converted to a deformation criterion. The load deflection behavior for an inelastic truss element, I believe, will be very simple and bi-linear.
I think you might be well-served in doing your post-processing external to SeismoStruct.
I'm still not sure what was going on with the joint constraint model you first proposed. If the post-processing you wish to do is more convenient for that model, then you could run both models and try to determine which number, load factor or base shear, is the appropriate one from the constrained model by comparison to the simpler model with inelastic trusses.
I think you are getting close to a solution though.
Kind regards,
huffte
Tim Huff
-
hani.akkari
- Posts: 18
- Joined: 20 Aug 2010, 22:00
Re: monitoring performance of steel braces
Dear Huffte,
thanks a lot for the help that you are proposing, 2 issues i want to clarify to you:
1- i didnt use the inelastic truss element because my final thesis CBF will be around 15 stories and five bays, can you imagine if i dont use the viual interface(prformance levels+ colors indicators), how much my life will be difficult to take every brace element and manipulate the results outside to get the targeted strain, meanwile if i use the frame element+release at the end , i can then assign to it performance levels (with colors ) so from the deflected shape i can know then immediately what brace elements has reached as performance level. i hope i'm clear in my choice.
2- thats why i have choosen to use the constraints in order to make pin the frame element ends(by using the equal degree of freedom for master and slave nodes). but the problem in this method is that there is a problem with the base shear , which come different than the product of load factor by the load pattern , i remodeled the same CBF using the inelastic truss elements , i got similar load factors as my constraints model but the base shear was direcly correct in the inelastic truss element, i think the software maybe confuse when slave and master nodes came on board.
as a conclusion i need to model my braces with frame elements to get benefit of the performance levels indicator but with making pin the ends of the frame element, my model with constraints was a try in this direction , but as soon as i got wrong base shear i have to change the constraints method and maybe use a different approach of releasing the ends, what about the link elements?
Best Regards,
Hani
thanks a lot for the help that you are proposing, 2 issues i want to clarify to you:
1- i didnt use the inelastic truss element because my final thesis CBF will be around 15 stories and five bays, can you imagine if i dont use the viual interface(prformance levels+ colors indicators), how much my life will be difficult to take every brace element and manipulate the results outside to get the targeted strain, meanwile if i use the frame element+release at the end , i can then assign to it performance levels (with colors ) so from the deflected shape i can know then immediately what brace elements has reached as performance level. i hope i'm clear in my choice.
2- thats why i have choosen to use the constraints in order to make pin the frame element ends(by using the equal degree of freedom for master and slave nodes). but the problem in this method is that there is a problem with the base shear , which come different than the product of load factor by the load pattern , i remodeled the same CBF using the inelastic truss elements , i got similar load factors as my constraints model but the base shear was direcly correct in the inelastic truss element, i think the software maybe confuse when slave and master nodes came on board.
as a conclusion i need to model my braces with frame elements to get benefit of the performance levels indicator but with making pin the ends of the frame element, my model with constraints was a try in this direction , but as soon as i got wrong base shear i have to change the constraints method and maybe use a different approach of releasing the ends, what about the link elements?
Best Regards,
Hani