Punching shear remains one of the least intuitive checks in concrete design. A flat slab supported directly on columns, with no beams, concentrates an enormous shear force into a small region around each column - and the failure, when it comes, is sudden and brittle, a cone of concrete punched clean through the slab. Two design approaches dominate, and they do not always agree.
Why it is hard to picture
Unlike beam shear, punching is a three-dimensional problem. The slab tries to fail along a roughly conical surface around the column, and the resistance depends on the concrete, the flexural reinforcement crossing that surface, and the geometry of the column itself. There is no clean shear-force diagram to read it off; the check is fundamentally about a perimeter and a stress on it.
The EC2 control-perimeter method
EC2 handles this with a control perimeter placed at a set distance from the column face - the basic perimeter sits at from the face, where is the effective depth. The applied shear is spread over this perimeter and compared to a resistance that scales with the concrete strength and, importantly, with the flexural reinforcement ratio to the one-third power. More tension steel through the column zone gives more punching resistance, because it controls the cracking that precedes the failure.
It is a fast, empirical, well-calibrated method - which is exactly why it is the everyday tool.
The fib Model Code approach
The fib Model Code uses the Critical Shear Crack Theory (CSCT), a more mechanical model. Instead of an empirical perimeter rule, it ties the punching resistance directly to the slab rotation at the column: the more the slab rotates, the wider the critical shear crack opens, and the less shear the concrete can transfer across it.
CSCT ties the resistance directly to the slab rotation - a more physical, if heavier, calculation than the control-perimeter rule.
It is more work, and it needs an estimate of the rotation, but it captures the underlying mechanics rather than fitting a curve to test data.
Where they diverge - and why it matters
For lightly reinforced slabs the two methods agree closely; you would pick whichever is more convenient. The interesting region is high reinforcement:
- For reinforcement ratios above roughly 1.5%, the two approaches start to diverge.
- That divergence can change a real design decision - specifically, whether a drop panel or column head is needed at all, or whether the slab can carry the punching demand on its own.
When the methods disagree on something that concrete, it pays to understand which assumptions each is leaning on, rather than picking the more generous answer by default.
Using this in practice
- EC2's control-perimeter method is the right default for routine flat slabs - fast, calibrated, and accepted.
- Reach for CSCT when the reinforcement is high and the control-perimeter result is the thing forcing an expensive drop panel; the more mechanical model may show the slab is adequate.
- Remember the flexural steel earns its keep twice - it resists the bending moments and boosts the punching resistance, which is why detailing the column-zone reinforcement well matters so much.
The EC2 punching provisions sit in the standards reference. The headline: punching is brittle and unforgiving, so the method you choose - and the reinforcement you detail through the column zone - deserves real attention, not a default.