Seismosoft | Earthquake Engineering Solutions

Hi,

I am Ajay, MS Civil Engineering student at IIT Madras, India. I am currently working on seismic response of Open Ground Storey (OGS) RC frames.

I am performing static pushover and dynamic time history analysis on OGS frames. For the analysis, 4 and 6 storey 2D RC frames are modeled with and without infill. Also variation regarding distribution of plasticity along the frame element is being considered. For lumped plasticity analysis I am using SAP2000 v17 whereas for distributed plasticity analysis Seismostruct v7 is being used.

Currently I am facing some issues in analysis.
1. Although geometric and material properties of models in both software are same, there is considerable difference in time periods indicated by two software. In SAP2000 infill is modeled as strut with pin joints at end as suggested by previous researchers, while in Seismostruct Crisafulli model is implemented using formulae suggested in manual. For bare frames time periods indicated by SAP2000 and Seismostruct are 0.534 sec and 0.421 sec respectively. For frames with Infill modelling they are 0.309 sec and 0.245 respectively. I have referred to previous research where comparison of modelling in both software was done and researchers did not find any large difference in time periods. I am unable to find a reason for difference in time periods.

2. In time history analysis for bare frame models, forces in frames members and deformations of top nodes are considerably higher in Seismostruct than SAP2000. While models with infill in SAP2000 show rise in ground storey column moments and shears, Seismostruct models show large reduction in ground storey column forces; which is not expected as per structural analysis. Since top storey move together as a single block, most of displacement takes place in ground storey columns only; hence large forces are expected in them compared to bare frame analysis. There is reduction in top node displacements in both software, but difference in Seismostruct models is very large. In SAP2000 max and min top node displacements for bare frame are 17 and -21 mm and for infilled frames are 11 and -6 mm. In Seismostruct they are 57 and -45 mm for bare frames and 2 and -2 mm for infilled frames. I am not able to figure out this anomaly.

If anybody has experienced same issues, please help me in this matter. I can send you analysis models via email.

Thanks in advance.

Hi ajaymore. As a first step I would make certain that material densities are identical in the two models and that all masses - whether input as specified masses or generated from load cases - are identical in the two models. Also, make certain that all rigid-end-joint sizes, modulus of elasticity, and all material properties are identical in the two models. These issues will hopefully help pinpoint the difference - there is either more mass in the SAP2000 model or more stiffness in the SeismoStruct model or a combination of both.

For time history, you need to make certain that all reinforcement details, confinement issues, P-Delta affect issues, and inelastic material properties are identical between the two models.

So, even though the overall geometry may be identical, there is still al lot to check for compatibility between two models.

Best of luck.