Web Openings In Steel Beams For Services

Services have to get through structure, and on a steel framed floor the choice is to run below the beams and lose ceiling height, or to pass through the webs and keep it.

Passing through is usually the right answer, and a hole cut in the wrong place removes capacity nobody checked.

What the web is doing

The flanges carry the bending. The web carries the shear and holds the flanges apart.

Which means a hole in the web is mostly a shear problem rather than a bending one, and the consequence depends entirely on where along the span it sits.

Near the supports, where shear is at its maximum and bending is at its minimum. The worst place for an opening, and the place services naturally want to go because it is closest to the riser.

At mid span, where bending is at its maximum and shear is close to zero. The best place for a web opening, and a hole there has a modest effect provided the flanges are untouched.

That single fact resolves most of these jobs. Move the penetration toward mid span and the problem largely disappears.

The failure modes

Web shear, because the remaining web above and below the hole has to carry the shear that the full web used to.

Vierendeel bending, which is the mechanism specific to web openings. The sections above and below the hole act as short members spanning across it, carrying their own local bending. The longer the opening, the more severe this becomes, which is why a long rectangular hole is far worse than a round one of the same depth.

Web buckling, where the slender strips left above and below a large opening buckle rather than yield.

Local yielding at the corners, which is why a rectangular opening has radiused corners and a square cut corner is a crack starter.

Deflection, because the beam is locally less stiff and the extra deflection accumulates across several openings.

The rules that keep it simple

Round beats rectangular. A circular hole generates far lower stress concentrations and avoids most of the Vierendeel problem.

Keep it in the middle third of the depth, so the strips above and below are equal and generous.

Never cut the flanges, or the fillet region joining the web to the flange. That region is where the section's capacity lives.

Stay away from point loads, from the supports, and from any connection or stiffener.

Space multiple openings far enough apart that they do not interact, which is usually at least the depth of the beam between them.

Keep the total depth of the hole to a modest fraction of the beam depth unless the opening is reinforced.

Reinforcing an opening

Where an opening has to be larger or closer to a support than the rules allow, it is reinforced.

Horizontal stiffener plates welded above and below the opening restore the local bending capacity. Vertical stiffeners either side restore the shear path. A full sleeve ring around a circular penetration does both.

All of which is welding onto an existing loaded member, which is a different proposition from fabricating it that way in the shop. Done on site it carries the hot works, access and sequencing constraints that make it expensive, and it is the argument for resolving the services coordination before fabrication rather than after.

Resolve it on the drawings, not on site

The pattern that causes problems is predictable. The steel is fabricated, the services are coordinated afterwards, and a duct arrives on site needing a hole where the beam cannot take one.

Which is why the penetration schedule belongs in the shop drawings, agreed between the services and the structure before anything is cut. That is part of what [steel connection design and shop drawings](/blog/steel-connection-design-and-shop-drawings) and [why engineering drawings must be buildable](/blog/why-engineering-drawings-must-be-buildable) exist to prevent.

A hole cut in the shop is accurate, reinforced if it needs to be, and protected against corrosion afterwards. A hole cut on site is none of those, and the bare steel at a flame cut edge is exactly where the problem in [corrosion protection for structural steel](/blog/corrosion-protection-for-structural-steel) starts.

Existing beams

Where a hole is needed in a beam already in service, the assessment is the same and the information is worse.

The section size has to be confirmed, the actual loads established, and the remaining capacity calculated before anything is marked out. On a building without drawings that is the exercise in [assessing an existing building with no drawings](/blog/assessing-an-existing-building-with-no-drawings).

Often the answer is a smaller hole in a different position, which the services can design around once they know. The expensive version is the hole cut first and the question asked afterwards.

ACSES provides structural engineering and structural assessment across Sydney. Talk to us about a project.