Inelastic Displacement Spectra Results

03-Unexpected behaviour/errors
Post Reply
BRAM
Posts: 2
Joined: 02 Aug 2016, 19:12

Inelastic Displacement Spectra Results

Post by BRAM »

Hi,

I would like to confirm if the results of the Inelastic Displacement Spectra corresponds to "yield-deformations response spectrum (u_y)" or to "pick deformations (u_m)" of inelastic system. (where ductility = u_m/u_y).

I have compared my calculations of inelastic spectra (according Chopra's book) and I find a difference with SEISMOSPECT exactly of target ductility's magnitude. So maybe SEISMOSPECT achieve u_m instead of u_y.

I find too a little difference between the Inelastic Acceleration Spectra, but it is not proportional to the ductility... same question. Is this acceleration associated to yield-deformation or to pick-deformation of inelastic system?

Thank you for any comment.
Bram
huffte
Posts: 889
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: Inelastic Displacement Spectra Results

Post by huffte »

Hi BRAM. It is not clear what you mean by "pick deformations". I will assume you mean "peak deformations".

First of all, which option for computing inelastic spectra did you choose? There are multiple options available in the settings, the most accurate fo which is the nonlinear response history analysis (SeismoSignal) method. In fact, it is the only accurate method for all cases. The others are simply attempts to establish simplified rules for inelastic spectra generation without resorting to inelastic analysis.

I am not sure which method you refer to in the Chopra book, but I suspect that this too is a simplified method useful when nonlinear response history analysis is unavailable.

The results are sensitive to the level of damping specified and to the post-yield stiffness ratio. Damping is taken equal to 5% for inelastic spectra in SeismoSpect and the post-yield stiffness is specified in the settings. If you want inelastic spectra for damping values other than 5% of critical, then SeismoSignal has this capability.

SeismoSpect is computing the maximum (peak) displacement at each period for a specified constant ductility. The SeismoSignal method option in SeismoSpect is the way to get the most accurate results. The simplified methods work well sometimes, not so well others.

Best of luck, BRAM.
Tim Huff
BRAM
Posts: 2
Joined: 02 Aug 2016, 19:12

Re: Inelastic Displacement Spectra Results

Post by BRAM »

Thanks, and sorry my language! :v

What I am doing is to compare an inelastic spectrum calculated from a SDOF model that I did in OpenSEES with the results of "SeismoSignal" method. This, with 5% damping and 0% post-yield stiffness (EPP), for EL CENTRO record. (The results of the model in OpenSEES were validated previously with the results of Chopra's book).

Now, with your response, I know the results are very similar (http://i.imgur.com/huPiioI.png).

I assume that for velocity and Acceleration Spectra it's also maximum response and not the "yield-response".

Thanks again,
bram
huffte
Posts: 889
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: Inelastic Displacement Spectra Results

Post by huffte »

Yes. Acceleration and velocity are maxima as well when computing inelastic spectra.

I have compared SeismoSpect to BiSpec and PRISM and obtained very similar results for each. IT would appear that OpenSEES is doing the task well too.
Tim Huff
Post Reply

Return to “03-Unexpected behaviour/errors”