Obtaining Correct Period for SAC LA Office Building

[rgrosa]

Hello, all. I'm an undergraduate research assistant at my university and am using OpenSees to model the SAC LA Office models (post-Northridge). I have read a few publications that use the 3-story model and all provide a fundamental period of around 1.01 - 1.1 seconds.

When I run my code, my model shows a period of ~.65 seconds. With a limited understanding of structural dynamics (I'm doing this research to teach myself some of the basics), I know that the error lies in the stiffness of my structure (I think I am 'over-constraining' something but I'm not 100% sure).

The code is long so I will try my best to explain EXACTLY how I defined my model:

I am modeling the SAC 3-story LA office building, post-Northridge design, with concentrated plasticity in ZLEs that are given the IMK material. There are 4 fixed supports and a 5th pinned support (for the PDelta column). Columns are defined as forceBeamColumns and connect directly to one another (no ZLE connections). Beams are modeled as elasticBeamColumns and are given large I values so that they act rigid. ZLEs connect the beams to the columns. These ZLEs are given rigid material tags in the translational directions and the IMK material in the appropriate rotational direction (I believe it is defined as 6).

The PDelta column stands on its own and transfers the PDelta effects through use of the equalDOF command to the main structure. Should the columns that make up this PDelta column have large E or I values to represent the aggregate effect of all the gravity columns as is done in the pushover analysis example? I currently am using the same columns that are defined in the actual moment frame. Also, for each floor I use the respective PDelta column node and equalDOF assign it to each node on that floor (only one node of each ZLE. Should I be constraining both?).

I believe that covers the important aspects of my model. I greatly appreciate any help or advice that can be provided! Thank you.

[fmk]

you might try ignoring the PDelta for now and see what the eigenvalues are (PDelta effect is probably negligable anyway for a 3 story building).
if that does not work change the model to elastic and recheck the eigenvalues.

and the code should not be long look at slide 29 of the last seminar we did.
http://opensees.berkeley.edu/wiki/index ... Techniques

and before you ask, an old version of the SteelWSections this is udes can be found here:
http://opensees.berkeley.edu/WebSVN/lis ... eesDays%2F