Structural Safety Factor

Structural safety factor

The global safety factor, FS = R / S, is just the resistance R (what the section or element can carry) divided by the solicitation S (the effect of the loads acting on it). Modern limit-state codes don't lean on one number. They break the margin into partial factors. On the resistance side you get γ_c = 1.4 for concrete and γ_s = 1.15 for steel (NBR 6118 §12.4); on the load side, γ_f = 1.4 for permanent actions and γ_f = 1.4–1.5 for the variable or accidental ones. Multiply it all out and the effective FS usually sits around 2.0–2.2.

Take R_k = 500 kN and S_k = 200 kN: the global FS is 500/200 = 2.5, which reads as safe. The ULS check is R_d = R_k/γ_s ≥ S_d = γ_f·S_k. Design for R_d ≥ S_d, and run the SLS checks (deflection, cracking, vibration) on their own. A generous FS tells you nothing about how the structure behaves in service.

Applications

Sizing elements under NBR 6118 (concrete) and NBR 8800 (steel). Checking columns, beams and slabs. Lining the result up against international codes such as Eurocode, ACI 318 and ASCE 7, where the partial factors come out differently. Gauging structural redundancy and robustness. And, on a retrofit, sizing up an existing building through its R/S ratio.

FAQ

Is FS = 2 always safe? Not on its own. ULS works with partial factors rather than a single FS, so "safe" really means R_d ≥ S_d once every γ is applied. A global FS of 2 papers over how much materials and loads can vary.

Why is γ higher for concrete than steel? Concrete's compressive strength scatters more than steel's, since batching and curing happen on site and are harder to pin down. That is what justifies γ_c = 1.4 against γ_s = 1.15. Steel comes off an industrial line with much tighter quality control.

What about ASCE 7 / LRFD? US codes go the LRFD route, with a resistance factor ϕ (< 1) and load factors such as the typical 1.2D + 1.6L. The logic mirrors NBR's γ_c/γ_s and γ_f, just flipped: ϕ multiplies R while γ multiplies S.