Shrinkage And Creep In Concrete Structures

Concrete does two things over time that steel does not. It shrinks as it dries, and it keeps deforming under a load that never changes.

Neither is a defect. Both are predictable, both are quantified in design, and both cause problems when the detailing ignores them.

Shrinkage

Fresh concrete contains far more water than the cement needs to hydrate. The surplus leaves over months and years, and the concrete gets smaller as it goes.

Typical drying shrinkage is in the order of several hundred microstrain, which on a twenty metre slab is a movement of a few millimetres. That sounds trivial until you consider that the slab is restrained at both ends by walls, columns and footings that have no intention of moving with it.

Restrained shrinkage puts the concrete into tension. Concrete has very little tensile capacity, so it cracks. The crack is the shrinkage finding somewhere to go.

Which means the design question is never how to stop shrinkage cracking. It is where the cracks will be and how wide they are allowed to get, which is the same argument as [crack width](/blog/crack-width-in-concrete-and-when-it-matters).

Plastic shrinkage is a separate and earlier problem, in the first hours while the concrete is still soft, driven by surface water evaporating faster than it bleeds. It produces short random surface cracks and it is a curing failure rather than a design one.

Creep

Load a concrete member and it deflects immediately. Leave the load there and it keeps deflecting, for years, at a decreasing rate.

The long-term creep deflection is commonly two to three times the immediate elastic deflection. On a member where the immediate deflection was acceptable, the final deflection may not be.

Creep depends on the age at first loading, the sustained load as a proportion of capacity, the member geometry and the humidity. Concrete loaded young creeps more, which is why propping cycles and early striking matter structurally and not just programmatically.

Why they matter together

Both are time-dependent, both are largely irreversible, and both act in the same direction on a typical floor: downward deflection and inward shortening.

The combination shows up as:

Cracking in masonry sitting on a concrete frame, because the frame shortens and the brickwork does not. This is the mechanism behind a large share of facade cracking in multi-storey buildings, and the reason horizontal movement joints exist in a masonry skin.

Doors and windows out of square where a transfer element deflected after the partitions went in.

Ponding on roofs and balconies, where a slab designed with a small fall lost it to long-term deflection.

Cracked tiling and failed waterproofing on a slab that moved after the wet area was finished.

Loss of prestress in post-tensioned members, where shortening relaxes the tendons. Allowed for in the design of [post-tensioned concrete](/blog/post-tensioned-concrete-when-and-why) and a real quantity, not a rounding error.

Differential column shortening in tall buildings, where a heavily loaded column shortens more than a lightly loaded one and the floors between them tilt.

How it is handled in design

Deflection calculated as long-term, not immediate, with the multipliers the standard sets. Covered in more detail under [deflection limits](/blog/deflection-limits-and-serviceability).

Precamber on long-span members, so the member is built up and deflects down to level.

Shrinkage reinforcement at a quantity that controls crack width rather than preventing cracks.

Joints that let the movement happen, which is the whole function of [expansion and control joints](/blog/expansion-and-control-joints-in-concrete).

Pour sequencing and closure strips, leaving a gap between large pours for weeks so most of the early shrinkage happens before the sections are connected.

Sequencing the finishes, meaning masonry, tiling and waterproofing installed after the bulk of the early movement rather than immediately behind the formwork.

What reduces it

A lower water content, which means a lower slump and more care in placing. Larger aggregate, because aggregate restrains the paste and the paste is what shrinks. Good curing, which slows the drying rather than stopping it. A higher strength grade, which tends to creep less under the same load ratio.

None of these eliminate it. The design allows for what remains, which is why a slab detailed by somebody who understands the numbers behaves and one detailed without them cracks on a predictable schedule.

ACSES provides structural engineering for residential and commercial projects across Sydney. Talk to us about a project.