Joints, limits, and what actually works

What actually reduces the cracking

Go back to w_k = s_r,max · ε and ask what each option changes.

Reduce the restraint R. This attacks ε directly and proportionally. Halving R halves the crack width. Achieved by casting in shorter bays, by leaving an infill strip to be filled a fortnight later, or by sliding membranes under a base slab. The most effective option, and it costs nothing in materials.

Reduce T₁. Also attacks ε directly. Achieved with a lower cement content, a cement replacement such as ggbs that hydrates more slowly, a lower placing temperature, or striking formwork later so the concrete cools gradually. A specification decision, not a calculation one.

Add reinforcement. Reduces s_r,max only, and through a 1/ρ term, so the returns diminish quickly. Going from 0.7% to 1.4% steel does not halve the crack width — it reduces the bar term only, and the cover term is untouched.

Use smaller bars. Reduces the bar term for the same area of steel, which is genuinely useful and costs nothing. Small bars closely spaced beat large bars widely spaced, exactly as in Module 6.

The reinforcement is the option a structural engineer reaches for first and the one that helps least.

Three kinds of joint

Construction joints are simply where one pour stops and the next begins. They are not intended to move. The reinforcement runs through, the surface is prepared, and a waterstop is usually cast in. Their position is a structural decision because it sets the bay size and therefore the restraint.

Movement joints are intended to move, and they come in two kinds. A contraction joint allows the concrete to shorten: the reinforcement is stopped, and the joint opens. A expansion joint allows movement both ways and includes a compressible filler. Both need a waterstop, and both are places where leakage is likely if the detailing is poor.

Induced joints are a compromise worth knowing about. A deliberate reduction of section — a fillet or a crack-inducing former — makes the concrete crack there rather than somewhere unpredictable. The crack still happens, but it happens at a place with a waterstop in it.

The design decision is between two strategies: provide enough steel to control cracking everywhere and use only construction joints (continuous construction), or provide movement joints at intervals and less steel between them (closed or partial contraction). Continuous construction uses more steel and fewer joints; joints are the more likely leakage path, so the trend has been towards continuous.

Autogenous healing

A fine crack in a water-retaining structure will often seal itself. Calcium hydroxide leaches out, meets carbon dioxide dissolved in the water, and precipitates calcium carbonate in the crack. Over weeks the crack closes.

This is a real effect and it is partly why the limits are as generous as they are. But it needs three things:

• a fine crack, below about 0.2 mm;
• still or slow-moving water, so the precipitate stays where it forms;
• time, typically weeks.

And it cannot be relied on for a crack that moves. A tank that fills and empties daily works its cracks open and closed, and any precipitate that forms is broken up again. Healing is a bonus for a crack that is essentially static, not a design mechanism for one that is not.