Hollow Section Connection Calculator

What it checks

Connecting a beam to a hollow section column is not the same problem as connecting to an open section. There is no flange to bolt through and no back face to reach, so the connection is normally a fin plate welded to the tube wall - and the tube wall itself becomes a failure mode that does not exist with an I-section column.

This calculator verifies the fin plate in gross shear, net shear, block tearing and bearing, the weld to the tube, the tube wall in local shear and punching shear, and chord face plastification to the CIDECT design rules. It handles circular (CHS), square (SHS) and rectangular (RHS) columns.

The tube wall, not the plate, is the new failure mode

Running the tool against the published example gives these values. Notice the difference in magnitude between the plate resistances and the wall resistances:

Fin plate, block tearing

516.6 kN

Fin plate, gross shear

525.1 kN

Fin plate, net shear

572.0 kN

Column wall, local shear

560.1 kN

Chord face plastification, N1,Rd

65.3 kN - GOVERNS the axial case

Chord face plastification, M1,Rd

18.28 kNm

Punching shear limit on plate thickness

7.88 mm

Column slenderness d/t

43.3

Chord face plastification at 65.3 kN is an order of magnitude below every plate resistance. That is the characteristic result for a hollow section connection: the tube wall bends locally under the plate long before anything in the plate itself yields, and it is why a connection detail that would be fine on an I-section column can fail badly on a tube of the same nominal capacity.

The punching limit of 7.88 mm is the other trap. If the fin plate is thicker than that, the plate can punch through the tube wall rather than the connection yielding - a brittle mode. A designer used to open sections will naturally reach for a thicker plate when a check fails, which here makes the connection worse rather than better.

Wall slenderness drives everything

The d/t ratio of the column - 43.3 in the example - is the parameter that controls chord face plastification. A thin-walled tube has very little local bending resistance, so the resistance falls off sharply as d/t rises. Doubling the wall thickness is far more effective than increasing the tube diameter at constant thickness.

The CIDECT rules that the chord face check follows have validity limits on d/t, on the plate-to-tube width ratio, and on the material. Outside those ranges the formulation is not applicable and the tool says so rather than extrapolating - a connection outside the validated range needs a different approach, typically a through-plate or a stiffened detail.

Formula reference

Punching limit on the plate

$t_{p,\max} = \dfrac{f_{u,c}\,d}{1.5\,f_{u,p}}$

Fin plate gross shear

$V_{Rd,g} = \dfrac{h_p t_p}{1.27}\cdot\dfrac{f_y}{\sqrt3\,\gamma_{M0}}$

Fin plate net shear

$V_{Rd,n} = A_{v,net}\dfrac{f_u}{\sqrt3\,\gamma_{M2}}$

Block tearing

$V_{eff,Rd} = \dfrac{f_u A_{nt}}{2\gamma_{M2}} + \dfrac{f_y A_{nv}}{\sqrt3\,\gamma_{M0}}$

Chord slenderness

$\gamma = \dfrac{d_0}{2\,t_0}$

Fillet weld throat

$a = \dfrac{s}{\sqrt2}$

Design weld strength

$f_{vw,d} = \dfrac{f_u}{\beta_w\,\gamma_{M2}}$

Assumptions and limits

Chord face plastification follows the CIDECT design rules, which are empirical and carry validity limits on the geometry ratios and on the steel grade. Outside those limits the expressions are not applicable - the tool flags the case rather than extrapolating, and a connection outside the range needs a through-plate, a stiffening ring or a different detail entirely.

The tube is assumed unfilled. A concrete-filled hollow section behaves very differently: the fill restrains the wall and chord face plastification largely disappears, but that is a composite design problem outside this tool. Chord axial and bending stress reduce the chord face resistance through a chord stress function - check whether the column loading in your case requires that reduction.

The tool checks the connection region. It does not check the column as a member, the beam as a member, fatigue, fire or seismic detailing.

FAQ