Storey drift limits in NBC 2020, and what they protect

The limits

NBC 2020 Division B, 4.1.8.13 sets interstorey deflection limits as a fraction of the storey height hs, and the fraction depends on what the building is for.

These are interstorey limits, not a limit on the total sway at the roof. Each storey is checked on its own, which is why a building that looks fine overall can fail at a single soft storey where the stiffness drops.

The amplification is the part that catches people

The deflections that go into this check are not the ones your analysis prints. Seismic design forces are reduced by the ductility and overstrength factors on the assumption that the structure yields and dissipates energy. The real displacement is larger than the elastic one computed under those reduced forces, so it has to be scaled back up.

Design deflection = elastic deflection x Rd Ro / IE

For a moderately ductile frame that multiplier is not a rounding adjustment. A system with RdRo around 5 and an importance factor of 1.0 turns a 12 mm computed interstorey deflection into 60 mm. Reporting the unamplified number is the most common way a drift check appears to pass when it does not.

Where the deflection is measured

Drift is not evaluated at a convenient node. The clause is concerned with the worst displacement anywhere on the storey, which means torsion has to be in the model.

• Displacements are taken including the torsional effects required elsewhere in 4.1.8, including accidental eccentricity.
• On an irregular or torsionally sensitive plan, the governing point is at the far corner, not the centre of mass.
• Both orthogonal directions are checked, and the combination rules for the two apply here as they do for forces.

Stability, and why drift and P-delta travel together

A drifted structure carries its gravity load off-plumb, and that eccentricity produces additional overturning that produces additional drift. NBC captures this with a stability factor computed per storey from the gravity load, the storey shear, the storey height and the amplified drift.

The practical relationship is simple: the checks are coupled. A storey that is close to its drift limit is also the storey most likely to require P-delta effects to be included, and including them makes the drift worse. If drift is marginal, resolve the stiffness rather than the arithmetic.

Wind drift is a different question

The 4.1.8.13 limits are seismic. Wind drift is governed by serviceability rather than by a prescriptive limit in Part 4, and the commonly used working figure of roughly 1/500 of the height is guidance from the Structural Commentaries, not a code requirement you can point at.

That distinction matters in two directions. You cannot cite 4.1.8.13 to justify a wind drift limit, and you cannot skip the wind serviceability conversation because the seismic drift passed. What the client actually cares about is usually neither number: it is whether the partitions crack, the cladding joints work, and the elevator rails stay within tolerance.

• Cladding and glazing have movement capacities set by their own manufacturers, and those can be tighter than anything in the code.
• Elevator guide rail alignment tolerances are frequently the controlling serviceability criterion in a tall building.
• Occupant comfort under wind is an acceleration problem, not a drift problem, and drift limits do not address it.

What to do when drift governs

In rough order of cost, from cheapest to most disruptive:

1. Check that the amplification and the torsional case were applied correctly before changing anything. A surprising share of drift failures are bookkeeping.
2. Redistribute stiffness rather than adding it. Moving a brace bay towards the perimeter buys more torsional stiffness than upsizing one in the core.
3. Deepen the members that contribute most to storey rotation. In a moment frame that is usually the beams, not the columns.
4. Change the system. If a moment frame is being sized by drift rather than by strength, a braced bay is doing the same job with less steel.
5. Revisit the ductility assumption. A higher Rd reduces the forces but increases the deflection multiplier, so it is not a free improvement.

Where CoLateral fits, and where it does not

CoLateral does not compute storey drift. It used to include a manual drift-ratio checker and that tool was retired, because entering storey heights and displacements by hand added nothing over an analysis program that reports interstorey drift directly and knows about the torsional case.

What the workspace is for is the part around the number: the frame behaviour you are reasoning about before the model exists, the seismic values the check depends on, and holding the drift result next to the assumption, the analysis run, and the review note that justify it. Six months later, that context is the thing nobody can reconstruct.