Seismosoft | Earthquake Engineering Solutions

Dear Madam/Sir,

After reading all the related papers given in the reference, the most suitable element class for my study is FB element. however it has a limitation of localization and research is being carried out to sort it out. FBPH is an alternative, but not suitable for my study. The difficulty which I am facing is related to the no of IPs; the results obtained shows a very high % of variation in the rotation value for 4 IPs, 5 IPs, 6 IPs. A difference of less than 10 % can be acceptable but in my case the difference of rotation value at a particular performance criteria is quite high as the no. of IPs are changed. However the moment values are approximately same even if the no. of IPs are changed which is satisfactory. What might be the possible error that I am making in my model? Please suggest.

P.S: going through the forum I found 'naserrahimzadeh' also encountered a similar problem, however the discussion never concluded.

thanks a lot.
regards,
Karabi

Hello,

I would like to add the output data obtained which is confusing me

for my model, for a particular performance criteria the value of rotation obtained is 0.0364, 0.024, 0.0172 and 0.0133 for 4, 5,6 and 7 IPs respectively. however the moment value obtained is 527kNm for all the IPs taken.

Please advise.

regards,
Karabi

The fact that the curvatures are more sensitive to the localisation effects is not unreasonable or unexpected. This is because the forces and moments have an upper limit (the capacity of the member) and in the highly inelastic range (as I expect you are in your analysis) small differences in the forces lead to much larger differences in the deformations, especially in the local level (e.g. curvatures).
When such 'sensitive' quantities are of importance, a sensinitivity study prior to the main analysis is problably required.
Seismosoft Support

Hello Seismosoft

thanks a lot for your time and valuable suggestion.

regards,
Karabi