Dead Load and Live Load

§ 01The weight that stays and the weight that wanders

Every structural calculation begins with a list of what the structure must carry. That list splits, almost immediately, into two kinds: load that is always there, and load that is not.

Dead load is the structure's own weight plus everything permanently attached to it — the steel frame, the concrete slab, the cladding panels, the fixed services running through the ceiling. It acts continuously, from the day the building is complete until the day it comes down. Because it is fixed and known, it can be calculated with reasonable confidence. Measure the geometry, multiply by the material density, and the number you get is close to the number that exists. Dead load does not surprise you.

Live load is everything else: people, furniture, stored goods, vehicles, snow on a roof, water ponding after a drain blocks. It comes and goes. It shifts position. It concentrates in ways that a designer has to anticipate without knowing exactly what will happen. A warehouse floor might sit empty for months and then receive a forklift carrying a full pallet, twice, on the same panel, at the same time. A theatre balcony holds a thin crowd on a Tuesday and a packed, standing, swaying audience on a Saturday. The structure must handle both without knowing which is coming.

The practical consequence is that dead and live loads are treated differently in analysis. Dead load is generally well-bounded — uncertainty is low, so the factor applied to it in a design check can reflect that. Live load carries more uncertainty about both magnitude and distribution, and the factor applied to it is correspondingly larger. The two are assessed separately and then combined according to rules that reflect how likely the worst case of each is to occur simultaneously. This is not arbitrary conservatism; it is a rational accounting for the different quality of information behind each number.

Distribution matters as much as magnitude. A live load applied to half a beam may produce a worse bending moment than the same total load spread evenly across the whole span. Standards for floor loading specify not just the intensity — load per unit area — but also a concentrated point load, placed wherever it produces the worst effect. Both have to be checked. The beam that passes the distributed check can still fail the point-load check, or vice versa.

One complication that sits between the two categories is superimposed dead load: finishes, raised floors, partition walls that are permanent in practice but sometimes altered during a building's life. A screed added during a refurbishment, or a partition repositioned, can shift the dead load without anyone revisiting the original structural calculations. This is one reason why the allowable stress in a code embeds assumptions about how loads will be characterised over a building's full life, not just at handover.

The distinction also matters for deflection. A beam under dead load deflects once and stays there; finishes are applied on top of the deflected shape, so that deflection is effectively invisible. Live load deflection happens repeatedly and is perceptible — it is what makes a floor feel springy, or makes a long-span roof visibly sag when snow accumulates. Codes limit live-load deflection separately from total deflection for exactly this reason: the number that matters to an occupant is the movement they can see, not the movement locked in before they arrived.

None of this means dead load is risk-free. A material that is denser than assumed, a topping slab poured thicker than drawn, a plant room added without structural review — dead load can exceed its calculated value too. The division into dead and live is a framework for organising uncertainty, not an assertion that one side of the ledger is already settled.

From the notes — How the two loads behave differently

Term	What it means here
Dead load	permanent, calculable, acts from completion; uncertainty is low, so design factors applied to it tend to be smaller
Live load	variable in magnitude and position; harder to bound, higher uncertainty, larger factors; must be checked in the worst-case arrangement, not just the average
Superimposed dead load	fixed in normal use but alterable over a building's life (screeds, partitions, finishes); sits between the two categories and is a known source of under-recognised risk