What is a partial safety factor, and why do we use one?

Neither the loads on a structure nor the strength of its materials are ever known with total certainty. The actual weight of a finished floor slab, the peak live load a room will ever really see, the true compressive strength of a specific batch of concrete — all of these are estimates with some natural scatter around them. A partial safety factor is a code-prescribed multiplier that accounts for that scatter, applied so that the design pushes toward the safe side on both ends: loads get multiplied up, and material strength gets multiplied down (or divided by a factor greater than one), before the two are compared.

The word "partial" is doing real work in that name. A single, blanket "factor of safety" applied once at the end of a calculation treats every source of uncertainty as if it were equally well known — which they are not. Splitting it into separate partial factors lets each one be calibrated to how confidently that specific quantity is actually known. Self-weight, for example, can usually be calculated quite precisely from drawn dimensions and known material densities, so its load factor is relatively small. Imposed (live) load — people, furniture, snow, whatever a space might realistically be loaded with over its life — is inherently harder to predict, so it gets a larger factor. On the resistance side, a factory-controlled structural steel section has less strength variability than site-mixed concrete, which is part of why the two materials' codes size their own strength-reduction factors differently.

Two systems, one underlying idea

Eurocode keeps this fully explicit: EN 1990 defines separate partial factors for permanent actions (γ_G) and variable actions (γ_Q), which are combined with the specific material-side factors set out in each material code (EN 1992 for concrete, EN 1993 for steel, and so on) to get design values on both sides of the check.

ACI 318 (working from ASCE 7 for the load side) reaches essentially the same place through a different structure: rather than a general combination formula with named partial factors, it gives a specific, enumerated list of load combination equations to check — the familiar 1.2D + 1.6L pattern and its variants — with the factors baked directly into each equation, and applies its own strength-reduction factors (φ) to the calculated capacity of each member type.

Both approaches are asking the same underlying question — "does the factored demand stay below the factored capacity, with enough margin to cover what we don't know for certain?" — they just organize the arithmetic differently. The Beam calculator works directly with the loads you enter to build shear and moment diagrams; the Load Combinations calculator is where those code-specific factors and combination rules actually get applied to a real set of actions (see the related tools below).