Seismosoft | Earthquake Engineering Solutions

i am doing a pushover analysis in a 3 storey bulding, and i face substantial numerical difficulties. Specifically I have used a large number of link element for the simoulation of the node. I applied any advice mentioned on the forum but still face significant numerical problems. any tip to overcome the numerical difficulties that may create the big number of link elements and rigid diaphragms that i use.

the analysis logo is:

LF= 1,00000, LF_incr= 1,00000 (Iter: 1 => Converg)


Variable Loading
----------------------------

Phase No. 1
Type of Control: Response Control
Controlled Node: 32C
Controlled Direction: y
Target Displ.: 0,3
Steps: 100
Disp_Incr= 0,00302, LF= 6,96893 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 10,16249 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 12,72270 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 14,70866 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 16,34822 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 17,66076 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 18,69171 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 19,51751 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 20,18404 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 20,75066 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 21,22569 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 21,64969 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 22,03581 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 22,32347 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= 22,59321 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= (Iter: 1 => fbd_Ite)
Disp_Incr= 0,00038, LF= 22,06367 (Iter: 1 => Converg)
Disp_Incr= 0,00075, LF= 21,80721 (Iter: 1 => Converg)
Disp_Incr= 0,00150, LF= 21,90998 (Iter: 1 => Converg)
Disp_Incr= 0,00038, LF= 21,49589 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= (Iter: 1 => fbd_Ite)
Disp_Incr= 0,00038, LF= 21,11491 (Iter: 1 => Converg)
Disp_Incr= 0,00075, LF= 20,96644 (Iter: 1 => Converg)
Disp_Incr= 0,00150, LF= 21,15116 (Iter: 1 => Converg)
Disp_Incr= 0,00038, LF= 20,81207 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= (Iter: 1 => fbd_Ite)
Disp_Incr= 0,00038, LF= 20,50255 (Iter: 1 => Converg)
Disp_Incr= 0,00075, LF= 20,40413 (Iter: 1 => Converg)
Disp_Incr= 0,00150, LF= 20,61566 (Iter: 1 => Converg)
Disp_Incr= 0,00038, LF= 20,32816 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= (Iter: 1 => fbd_Ite)
Disp_Incr= 0,00038, LF= 20,06331 (Iter: 1 => Converg)
Disp_Incr= 0,00075, LF= 19,99426 (Iter: 1 => Converg)
Disp_Incr= 0,00150, LF= (Iter: 1 => fbd_Ite)
Disp_Incr= 0,00019, LF= 19,67312 (Iter: 1 => Converg)
Disp_Incr= 0,00038, LF= 19,46768 (Iter: 1 => Converg)
Disp_Incr= 0,00075, LF= 19,45313 (Iter: 1 => Converg)
Disp_Incr= 0,00056, LF= 19,34140 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= (Iter: 1 => fbd_Ite)
Disp_Incr= 0,00038, LF= 19,15801 (Iter: 1 => Converg)
Disp_Incr= 0,00075, LF= 19,14784 (Iter: 1 => Converg)
Disp_Incr= 0,00150, LF= (Iter: 1 => fbd_Ite)
Disp_Incr= 0,00019, LF= 18,88535 (Iter: 1 => Converg)
Disp_Incr= 0,00038, LF= 18,73474 (Iter: 1 => Converg)
Disp_Incr= 0,00075, LF= 18,76680 (Iter: 1 => Converg)
Disp_Incr= 0,00056, LF= 18,68735 (Iter: 1 => Converg)
Disp_Incr= 0,00300, LF= (Iter: 1 => fbd_Ite)
Disp_Incr= 0,00038, LF= 18,54782 (Iter: 1 => Converg)
Disp_Incr= 0,00075, LF= 18,56933 (Iter: 1 => Converg)

What, in particular, causes you to doubt your results manousos? In looking over your log I see that the structure has reached a displacement of about 0.15 - one-half of the target - and that the structure has already gone past the peak load. So perhaps the target displacement is simply unreachable structurally?

Thank you very much for your answer Huffte. it is a logical answer. I have one more question. i have define the target diplacement as the three percent of the total height of the structure(3%htotal). what criteria should i choose to define the target diplacement??

Well, I believe that the answer to that question is up to you manousos.

Do you wish to achieve 3% of the story height? Then your analysis would indicate that you need to modify the structural details - reinforcement and/or cross sections - until you achieve the 3%.

Are you simply wishing to evaluate the structure as is? Then you should set the target displacement slightly beyond the failure limit state you wish to predict. This will require some iterations, but you do not have to get exactly the target you want, just beyond it and with a fine enough displacement increment to capture all the behavior you wish to track and report.

Best of luck, manousos.

Dear manousos,

A target displacement twice the displacement at the yield of the structure is typically enough to get all the required structural information that you need.

SeismoSoft Support

I would offer this qualifier to the statement that "twice the yield" is a reasonable target displacement. Many structures are permitted to undergo much larger displacements during strong ground shaking. For example, bridge structures in the US codes are permitted to respond at displacement ductility values as high as 6 (permissible displacement equal to 6 times the yield displacement) under certain conditions. So I would encourage users to know their goal when setting target displacements.

Hi manousos,

Huffte has a valid point. Some structures do have a displacement capacity that goes well beyond twice the yield displacement. In any case, it would be immediate for you to realise, as the analysis progresses, if you aimed for a too large a target displacement, or the inverse.

Another thumb rule can be to consider a target displacement of 2% or 3% the building height, but, again, it will be up to you, at the end of the analysis, to decide if you have pushed the structure enough or not.

It's all good fun!

Rui

Good fun indeed Rui. All made possible by the excellent SeismoSoft team!

Thanks Tim and Rui,
On second thought, you are both very correct; settings the target displacement equal to twice the displacement at yield is indeed not enough for some types of structures. (it seems that lately we are very influenced by the behaviour of old, lightly reinforced buildings [:)] )

Warm Regards,
SeismoSoft Support