IDA - direction of loading

[morteza.tahmasbi]

Dear support team
I model a 2D concrete frame. i want to obtain DM-IM curve for X direction. I assign time history acceleration to base nodes in X direction, should I assign the same acceleration in Y direction too?
Best regards

[z.gronti]

Dear Morteza,

Could you be more specific on which curve you want to obtain? I am not sure that i understand the DM-IM curve.

[morteza.tahmasbi]

2 distinct cases:
in first case maximum story drift is selected as DM (Damage Measure) and Spectarl acceleration of the first mode for 5% damping ratio is selected as IM (Intensity Measure). in the second case DM is the same as previous case but PGA is selected as IM. the input motion is Landers record and is scaled from 0.1 to 5.

[z.gronti]

You shouldn't apply in Y direction the same time history acceleration with the one in X direction. You should either apply one time history acceleration in X direction, or two different time histories accelerations in X and Y directions.

[ruipinho]

Hi morteza.tahmasbi,

If you have modelled a 2D frame in direction X, applying an accelerogram in the Y direction does not seem logical to me, given that there is no structural system in such direction. But maybe I have misunderstood your question?..

Best,

Rui

[z.gronti]

Dr. Rui Pinho is right. I have passed by the information regarding the 2D frame. My previous post refers to 3D frame models.

[seismosoft]

Also it should be noted that in the 3D case the two 'different' time histories in the X and Y directions should refer to the acceleration measurements in the two orthogonal directions X & Y on the same location and the same seismic excitation. One cannot apply any two records in the two horizontal directions.

[morteza.tahmasbi]

Dear support team
I'm so sorry for typing mistake, I meant I'm modelling a 3D concrete frame. I applied two orthogonal acceleration time histories in X and Y direction. I wanted to know if I applied record just in one direction for a 3D model would it yield to wrong results?
and if I take effective duration of acceleration time history instead of full record time (for computational simplicity and time saving) how are the results affected?
meanwhile the curve I'm obtaining after process seems to have a hardening behavior though it's expected that the curve tangent decrease when record intensity increases (IM-DM curve). what could be the reason of such a behavior?
Best regards

[huffte]

It depends on the structure you are modeling. If you have moment frame in both directions with inelastic members for columns, as an example, then applying the load in one direction only could certainly have a significant impact on results. And the duration issue depends on what you wish to capture. For example, if you want to capture residual displacements, then not only should you use the entire accelerogram, but you should probably pad it with several seconds worth of zeroes at the end to get an accurate value. So, in summary, give some thought to what type of model you have and which results are important to you in making decisions like these. These are good questions, morteza.tahmasbi. Best of luck.