Rekan Daho’s Post

View profile for Rekan Daho, graphic

NHRE Master‘s Degree @ Bauhaus Universität Weimar

Hind Hammada: Why does the ACSE code require scaling dynamic forces for seismic design, and what is the specific scaling factor? Almost all dynamic analysis results are less than of static ones (such as ELF). So the values like Base Shear tend to be smaller in dynamic analysis than ELF results. In ASCE7 you scale up to 100% ELF so you get little bump if things go your way. I agree in principle, though, that the choice to go RSA/NLA is usually driven more by having identified the need to capture complex dynamic behavior than by a desire to save on construction costs. ASCE 7-22 12.9.1.4 #RekanDaho

  • graphical user interface, application
ANAS DAWAS

Structural Consultant | Extensive Experience in ACI & ASCE Codes

7mo

First of all, we have to know that we have two different scaling; the scaling of forces and the scaling of drifts, so the MRSA forces used to design the elements may be differ from forces used to check drift (as I said; may be) The scaling of forces starts from initial scaling ( gI/R) to take the effect of importance factor and response modification factor , then we have to do scaling up 100% of ELF (scaling down is not permitted) because the computed fundamental period may be the result of an overly flexible (incorrect) analytical model (In other words; flexible building may results in very low base shear). Scaling of Drifts Displacements from the modal response spectrum are only scaled to the ELF base shear where Vt is less than CsW and Cs is determined based on Equation (12.8-7). For all other situations, the displacements need not be scaled, because the use of an overly flexible model will result in conservative estimates of displacement that need not be further scaled. So finally we have; ELF (Forces) for strength design ELF (Forces) for drift check MRSA (Forces) for strength design MRSA (Forces) for drift check

To view or add a comment, sign in

Explore topics