Applying Vertical Seismic Effects in SeismoStruct Nonlinear Analysis: A Practical Workaround
Posted: 11 Mar 2026, 17:58
Hello everyone,
I'd like to share a workaround for applying vertical seismic effects in SeismoStruct's gravity load step.
Many codes require this. For example, TBDY 2018 defines the nonlinear combination as:
G + Qe + 0.2S + Ed(H) + 0.3×Ed(Z)
where G=dead load, Qe=effective live load, S=snow, Ed(H)=horizontal and Ed(Z)=vertical seismic effect. Here 0.3×Ed(Z) ≈ 0.2×SDS×G.
Physically, this is equivalent to increasing gravitational acceleration to:
g' = (1 + 0.2×SDS) × 9.81 m/s²
In Project Settings > Gravity and Mass, simply change the g value. With "Loads derived from Masses" enabled, all masses are multiplied by g', increasing vertical forces in Step 1. The mass matrix is unaffected, so periods and mode shapes remain correct.
IMPORTANT: Enter all loads via the Modeler while g=9.81 first, because the Modeler uses the current g for load-to-mass conversion. Then change g before running the analysis. Reset afterwards.
LIMITATION: This also amplifies live load contributions (n×Q) in mass. The overestimation is small (~6% for typical values) and conservative.
A load type classification (Dead/Live/Snow) in future versions would allow applying vertical seismic effects only to dead loads.
Hope this helps!
Dr. Abdurrahman Sahin
I'd like to share a workaround for applying vertical seismic effects in SeismoStruct's gravity load step.
Many codes require this. For example, TBDY 2018 defines the nonlinear combination as:
G + Qe + 0.2S + Ed(H) + 0.3×Ed(Z)
where G=dead load, Qe=effective live load, S=snow, Ed(H)=horizontal and Ed(Z)=vertical seismic effect. Here 0.3×Ed(Z) ≈ 0.2×SDS×G.
Physically, this is equivalent to increasing gravitational acceleration to:
g' = (1 + 0.2×SDS) × 9.81 m/s²
In Project Settings > Gravity and Mass, simply change the g value. With "Loads derived from Masses" enabled, all masses are multiplied by g', increasing vertical forces in Step 1. The mass matrix is unaffected, so periods and mode shapes remain correct.
IMPORTANT: Enter all loads via the Modeler while g=9.81 first, because the Modeler uses the current g for load-to-mass conversion. Then change g before running the analysis. Reset afterwards.
LIMITATION: This also amplifies live load contributions (n×Q) in mass. The overestimation is small (~6% for typical values) and conservative.
A load type classification (Dead/Live/Snow) in future versions would allow applying vertical seismic effects only to dead loads.
Hope this helps!
Dr. Abdurrahman Sahin