Seismosoft | Earthquake Engineering Solutions

Hello

I am performing an eigenvalue analysis on a 5 span bridge with Seismo Struct 5.2.2. My model only contains the superstructure, built with “elfrm” elements, which is held by “bl_sym” link elements (I want to do a pushover later) according to the stiffnesses of the piers and abutments. I perform the eigenvalue analysis with Seismo Struct and Statik 6 for the exact same model and get different results (I am looking here especially at the slowest longitudinal mode). Seismo Struct gives me a period of T=0.60s and Statik-6 T=0.64s. Because of the EA and the length of the bridge (337m), there is a significant longitudinal wave in the superstructure. So to investigate the problem I set the EA very high in both models and tried again. With the rigid superstructure I can calculate the right period by hand very easily and get T=0.524s. Statik-6 gives me exact this period with the rigid superstructure whereas Seismo Struct calculates T=0.488s. I then changed the “bl_sym” elements in Seismo Struct to “lin_sym” elements with the same stiffness and so I get also exact the T=0.524s as by hand and with Statik-6.
Is it possible that there is somewhere a problem using “bl_sym” elements in a system and performing an eigenvalue analysis? In my opinion there should be no difference in the result of the eigenvalue analysis using either “bl_sym” or “lin_sym” since I gave them exactly the same stiffness. And if there is a difference caused by the calculation of the program it should in my eyes be rather softer than stiffer with the bl springs.
Have you come accross a similar problem in the past?

Bets Regrds,

Pascal

I am just wondering if maybe you included rotational stiffness about the transverse bridge axis in your link definition, Bridge87? Depending upon the actual boundary support conditions, a value of zero might be appropriate for this DOF.

This could artificially prevent a rotation which may actually be able to occur in the structure.

Just a thought. You've probably already thought of that, but it is the first thing that comes to mind.

Also, did you try the bl_kin element and get the same answer as with the bl_sym element? They behave slightly differently but should be identical for a simple eigenvalue analysis it would seem. Just another thought.

Best of luck.

Hello huffte

Thank you for your inputs. The rotational stiffness of all the links is set to zero for the transversal directions (horizontal and vertical) to the bridge. I only hold it with a rotation restraint in x direction (longitudinal to the superstructure) to prevent a rotation mode.

About the bl_kin elements I already thought yesterday evening, but I was too tired to try. I tried now and got the right result of T=0.524s. So it works with the lin_sym and the bl_kin links, but not with the bl_sym.

Best Regards