UK Steel Truss Design — EN 1993-1-8 Welded Joints in Hollow Sections

Welded hollow section trusses are a structurally efficient and architecturally expressive framing solution for long-span UK roofs, bridge footbridges, entrance canopies, and exposed architectural steelwork. Unlike bolted trusses with gusset plates, hollow section trusses use direct welded connections where the brace ends are profile-cut and fillet-welded directly to the chord face. Joint design is governed by BS EN 1993-1-8 Chapter 7 "Hollow Section Joints", which provides empirical resistance formulas derived from extensive CIDECT research programmes. This reference covers joint classification (gap vs overlap), the six failure modes of welded hollow section joints, chord face plastification, brace effective width, chord shear interaction, and a fully worked example for a 25 m span Warren roof truss.

UK Hollow Section Truss Configurations

Warren Truss (Standard UK Roof Truss)

The Warren truss with CHS (Circular Hollow Section) chords and CHS braces is the most common UK configuration. Equilateral or near-equilateral triangles formed by alternating diagonal braces carry both tension and compression. Typical UK parameters:

• Span-to-depth ratio: 12-18 for roof trusses (25 m span => 1.5-2.1 m truss depth)
• Panel length: equal to truss depth (45-60 degree brace-to-chord angle)
• Top chord: continuous CHS with site-welded or bolted splice at apex
• Bottom chord: continuous CHS, may be fabricated from shorter lengths for transport
• Braces: CHS profile-cut to match chord profile, fillet-welded all around

Pratt Truss

Less common in UK hollow section construction because the vertical web members create joint congestion and the alternating tension-compression diagonals produce lower joint efficiency than the Warren configuration. Used primarily where vertical glazing bars or secondary framing align with the panel points.

Vierendeel Truss

A truss with no diagonal members — the chords and vertical posts form rectangular panels that resist shear through bending of the chord and post members. Joints are full-strength moment connections. Used in UK buildings where diagonal braces would obstruct windows, doors, or service zones. Significantly heavier than Warren or Pratt trusses of the same span (typically 1.5-2.5 times the steel weight).

Joint Classification — Gap vs Overlap Joints

EN 1993-1-8 Clause 7.1.2 classifies K- and N-joints as:

Gap Joints

The toes of adjacent braces on the same chord face are separated by a gap g >= t1 + t2, where t1 and t2 are the brace wall thicknesses. Gap joints are the default UK configuration because:

• The gap provides weld access around the full perimeter of each brace
• The chord face between the braces is unrestrained and free to deform (chord face plastification is the governing failure mode)
• Fabrication is simpler — each brace can be profile-cut and welded independently

Minimum gap: g >= t1 + t2 for weld access. UK practice typically specifies 20-30 mm for trusses with braces up to 150 mm diameter. For larger braces, the gap should be scaled proportionally.

Overlap Joints

One brace overlaps the other on the chord face, with the overlapping brace welded to both the chord and the overlapped brace. Overlap joints are used when:

• The gap would be negative (braces would intersect before reaching the chord face)
• Higher joint capacity is required (overlap joints mobilise the brace wall in shear, which is typically stronger than chord face bending)
• The truss depth is constrained and the panel length cannot be increased to create a gap

The overlap must be at least 25% of the overlapped brace width and the overlapping brace must be welded to the overlapped brace. Overlap joints have higher fabrication cost (3 weld profiles instead of 2) and are less common in UK practice.

Failure Modes for Welded Hollow Section Joints

EN 1993-1-8 Table 7.1 defines six failure modes for welded CHS and RHS joints:

Mode A: Chord Face Plastification (Governing for Gap Joints)

The chord face yields in bending under the brace axial load, forming a plastic mechanism in the chord wall. This is the most common governing mode for gap joints with typical UK truss geometry.

Resistance for CHS K gap joints (brace in compression):

N_i,Rd = (f_y0 x t0^2 x k_g x k_p x k_n) / (sin(theta_i) x gamma_M5) x f(beta, gamma, theta)

Where:

• beta = d1/d0 (brace-to-chord diameter ratio)
• gamma = d0/(2 x t0) (chord slenderness)
• k_g accounts for the gap size g
• k_p accounts for chord preload from adjacent members
• k_n accounts for chord stress (axial stress in the chord reduces joint capacity)

Mode B: Chord Shear (Governing for High Shear, Deep Trusses)

The chord cross-section yields in shear between adjacent braces due to the vertical components of brace forces. For a Warren truss with 45-degree diagonals and equal brace forces, the chord shear V_0 is approximately 0.7 x V_panel (where V_panel is the vertical shear in the truss panel). Chord shear interaction reduces the available chord face plastification resistance.

Mode C: Chord Side Wall Buckling (RHS Joints Only)

For RHS joints where the brace width approaches the chord width (beta approaching 1.0), the chord side walls buckle under the compressive brace force. This mode is specific to RHS joints and does not occur in CHS joints. The chord web acts as a strut between the chord flanges.

Mode D: Brace Effective Width (Brace Failure)

When the brace width is small relative to the chord width (low beta), the brace cross-section is not fully effective in transferring the brace force to the chord face. Only a portion of the brace perimeter (the effective width) participates in load transfer. The effective width depends on beta and the brace slenderness.

Mode E: Chord Punching Shear (High Beta, Thin Chord Wall)

For high beta values (brace nearly as wide as the chord) with a thin chord wall, the brace can punch through the chord face in shear. The punching shear perimeter is the projected contact area of the brace on the chord face. This mode rarely governs in UK trusses with standard tube sizes.

Mode F: Brace or Chord Failure by Yielding or Buckling

The members themselves fail in tension yielding, compression buckling, or combined axial force and bending. This is NOT a joint failure mode per se but must be verified alongside the joint checks.

Joint Resistance Formulas for UK CHS Trusses

CHS K Gap Joint — Chord Face Plastification

For a gap K joint in CHS with the brace in compression:

N_i,Rd = (f_y0 x t0^2 / sin(theta_i)) x (1.8 + 10.2 x d1/d0) / (gamma_M5) x k_g x k_p

Where:

• f_y0 = chord yield strength (N/mm^2)
• t0 = chord wall thickness (mm)
• theta_i = brace-to-chord angle
• d1/d0 = brace-to-chord diameter ratio (must be 0.2 <= beta <= 1.0)
• k_g = gamma^0.2 x (1 + (0.024 x gamma^1.2) / (exp(0.5 x g/t0 - 1.33) + 1))
  • The gap function k_g decreases as the gap increases, reflecting the reduced chord face restraint.
• k_p = chord preload factor = 1.0 when chord axial stress ratio n < 0.4 (compression) or n < 0.0 (tension)
  • For higher chord stress: k_p = 1 + 0.3 x n - 0.3 x n^2 (where n = N_0,Ed / N_pl,0,Rd)

CHS K Gap Joint — Chord Shear

V_pl,0,Rd = (f_y0 / sqrt(3)) x A_v,0 / gamma_M5

Where A_v,0 = 2 x A_0 / pi (the shear area of a circular hollow section).

The chord shear utilisation V_0,Ed / V_pl,0,Rd must be less than 1.0. Partial interaction with chord face plastification is accounted for by reducing the chord face capacity when V_0,Ed / V_pl,0,Rd > 0.5.

CHS K Gap Joint — Brace Effective Width

N_i,Rd = f_yi x t_i x (2 x h_i - 4 x t_i + 2 x b_e,ov) / gamma_M5

Where the effective width b_e = (10 x f_y0 x t0) / (f_yi x t_i) x d_i (limited to d_i). This check ensures that the full brace cross-section can deliver the applied force to the chord face.

RHS K Gap Joint (Rectangular Hollow Sections)

The RHS joint formulas are similar but account for the flat chord face geometry, which produces more uniform plastification than the curved CHS face. RHS joints are less common in UK trusses because:

• CHS provides a more uniform stress distribution at the joint
• CHS has no weak-axis issues for the chord in compression
• CHS offers lower wind drag for exposed roof trusses
• RHS is used where bolted connections are required at panel points for transport or site assembly

Chord Member Design in Trusses

The chord members in a truss are subject to combined axial force and bending:

• Top chord: Continuous over panel points. Primary action is compression (from global truss bending), with secondary bending from purlin loads applied between panel points. The chord is designed as a beam-column per EN 1993-1-1 Clause 6.3.3.
• Bottom chord: Continuous or spliced at panel points. Primary action is tension (from global bending), with secondary compression under wind uplift reversal. The bottom chord typically has a smaller section than the top chord because tension members are not subject to buckling.

For a Warren truss, the axial force in the top chord at mid-span is approximately:

N_c = M_span / h_truss

Where M_span is the mid-span bending moment on the simply supported truss, and h_truss is the truss depth (centre-to-centre of chords). For a 25 m span truss at 1.5 m depth carrying 8 kN/m ULS:

N_c = (8 x 25^2 / 8) / 1.5 = (625) / 1.5 = 417 kN compression

UK Cost Data for Hollow Section Trusses

Approximate fabricated-and-erected costs for UK steel trusses (2026 prices):

Hollow section trusses carry a 15-25% cost premium over bolted angle trusses, but this is often offset by reduced corrosion protection costs (40% less surface area), lower maintenance, and architectural value for exposed steelwork.

Worked Example — 25 m Warren Roof Truss

A sports hall in Manchester requires a 25 m clear-span roof truss at 6.0 m centres. Design a CHS Warren truss for the following loads:

Design data:

• Span: 25 m
• Truss spacing: 6.0 m
• Truss depth: 1.8 m (span/depth = 13.9)
• Panel length: 1.8 m (equilateral triangles, brace angle = 45 deg)
• Number of panels: 14 (13 panel points + 2 supports)
• Steel grade: S355J2H to EN 10210

Step 1 -- Loading:

• Dead: roof cladding + purlins + services = 0.35 kN/m^2 x 6.0 m = 2.10 kN/m
• Snow: 0.50 kN/m^2 (Manchester ground snow) x 0.8 (shape coefficient) x 6.0 m = 2.40 kN/m
• Total ULS: 1.35 x 2.10 + 1.50 x 2.40 = 2.835 + 3.600 = 6.435 kN/m

Step 2 -- Global analysis:

• Maximum shear at support: V = 6.435 x 25 / 2 = 80.4 kN
• Maximum bending at mid-span: M = 6.435 x 25^2 / 8 = 502.7 kN.m
• Maximum top chord force: N_c = M / h = 502.7 / 1.8 = 279.3 kN (compression)
• Maximum bottom chord force: N_t = 502.7 / 1.8 = 279.3 kN (tension)
• Maximum diagonal force (45 deg): N_diag = V / (2 x sin(45)) = 80.4 / (2 x 0.707) = 56.9 kN per diagonal at support

Step 3 -- Member selection:

• Top chord: CHS 168.3 x 6.3 (A = 32.1 cm^2, i = 5.73 cm, Class 1 in compression)
  • Buckling length: 1.8 m (panel length, pinned at joints)
  • slenderness lambda_bar = 1.80 / (0.0573 x 76.4) = 0.41
  • Reduction factor chi = 0.93 (buckling curve a for CHS hot-finished)
  • N_b,Rd = 0.93 x 32.1 x 100 x 355 / (1.0 x 1000) = 1,061 kN > 279.3 kN -- OK
• Bottom chord: CHS 139.7 x 5.0 (A = 21.2 cm^2)
  • N_pl,Rd = 21.2 x 100 x 355 / (1.0 x 1000) = 753 kN > 279.3 kN -- OK
• Diagonals: CHS 88.9 x 4.0 (A = 10.7 cm^2)
  • Buckling length: 2.55 m (brace length for 1.8 m panel at 45 deg)
  • N_b,Rd = 0.72 x 10.7 x 100 x 355 / (1.0 x 1000) = 274 kN > 56.9 kN -- OK (compression braces at support)
  • N_pl,Rd = 10.7 x 100 x 355 / 1000 = 380 kN > 56.9 kN -- OK (tension braces)

Step 4 -- Joint verification at support panel:

The support panel has the highest brace forces (56.9 kN each from adjacent diagonals, one tension, one compression). Check as CHS K gap joint:

• Chord: 168.3 x 6.3, d0/t0 = 168.3/6.3 = 26.7
• Braces: 88.9 x 4.0, d1 = d2 = 88.9 mm, t1 = t2 = 4.0 mm
• Gap g = 25 mm (selected for weld access, g/t0 = 25/6.3 = 3.97)
• beta = d1/d0 = 88.9/168.3 = 0.528 (within 0.2-1.0 range)
• gamma = d0/(2t0) = 168.3/(2x6.3) = 13.4
• theta_1 = theta_2 = 45 degrees

Chord face plastification resistance (compression brace):

k_g = gamma^0.2 x (1 + (0.024 x gamma^1.2) / (exp(0.5 x g/t0 - 1.33) + 1)) = 13.4^0.2 x (1 + (0.024 x 13.4^1.2) / (exp(0.5 x 3.97 - 1.33) + 1)) = 1.68 x (1 + (0.024 x 22.3) / (exp(0.655) + 1)) = 1.68 x (1 + 0.535 / (1.925 + 1)) = 1.68 x (1 + 0.183) = 1.99

Chord preload: n = N_0,Ed / N_pl,0,Rd = 279.3 / 1,061 = 0.263 (compression, < 0.4) k_p = 1 + 0.3 x 0.263 - 0.3 x 0.263^2 = 1 + 0.079 - 0.021 = 1.058

N_1,Rd = (355 x 6.3^2 / sin(45 deg)) x (1.8 + 10.2 x 0.528) / 1.00 x 1.99 x 1.058 = (355 x 39.69 / 0.707) x (1.8 + 5.39) x 1.99 x 1.058 / 1000 = 19,930 x 7.19 x 1.99 x 1.058 / 1000 = 301.7 kN > 56.9 kN -- OK (utilisation 0.19)

The joint has substantial reserve capacity. The brace effective width and chord shear checks are also satisfied by inspection for this low utilisation.

Step 5 -- Deflection check:

Maximum vertical deflection at mid-span under characteristic loading (dead + snow, unfactored):

w_max = (5 x 4.50 x 25^4) / (384 x 210,000 x I_truss,equivalent)

Approximate I_truss for a Warren truss using parallel axis theorem: I_truss = 2 x A_chord x (h/2)^2 = 2 x 3210 x (900)^2 = 5.20 x 10^9 mm^4

w_max = (5 x 4.50 x 25^4 x 10^12) / (384 x 210,000 x 5.20 x 10^9) = 33.4 mm

Span/deflection = 25,000 / 33.4 = 749 > 250 -- OK for roof truss (recommended limit span/250 per SCI guidance).

Design Resources

• UK Connection Design Guide — Complete EN 1993-1-8 joint reference
• UK Hollow Section Connections — CHS and RHS connection design
• UK Portal Frame Design — Portal frames with trussed rafters
• UK Steel Warehouse Design — Warehouse framing with roof trusses
• UK Weld Design Guide — Fillet weld capacity for truss joints
• UK Steel Framing Cost Guide — Cost data for UK truss fabrication
• UK Beam Design Guide — Truss chord design as beam-columns

Frequently Asked Questions

Why use hollow sections for UK steel trusses rather than open sections with gusset plates?

Welded hollow section trusses eliminate gusset plates entirely, reducing fabrication cost by 15-25% compared to bolted angle trusses. Closed sections provide inherent torsional stiffness — critical for compression chords where lateral-torsional buckling of open sections (angles, channels) would reduce capacity. The clean, closed profile provides an architectural appearance valued for exposed steelwork in sports halls, atria, and entrance canopies. The approximately 40% lower external surface area reduces corrosion protection costs (painting or galvanising) and ongoing maintenance. The primary limitation is lower joint capacity: CHS joints governed by chord face plastification require larger chord sections than a comparable gusset-plated truss, and heavily loaded trusses (e.g., transfer trusses supporting multi-storey columns) may exceed the validated parameter range for EN 1993-1-8 joint formulas.

What is the minimum gap between braces in a UK K joint?

EN 1993-1-8 Clause 7.1.2 requires a minimum gap g >= t1 + t2 between the toes of adjacent braces on the chord face. This is a fabrication requirement — the gap must be sufficient for the welder to access the full perimeter of each brace with the welding torch. For UK trusses with typical CHS braces up to 150 mm diameter, the gap is 20-30 mm. If the brace angle is shallow (theta < 30 degrees), the projected gap on the chord face is larger than the perpendicular gap and must be checked separately. If the gap is less than t1 + t2, the joint must be classified as an overlap joint and designed using the overlap joint formulas in EN 1993-1-8 Table 7.1, which involve different (and generally more complex) resistance calculations.

Do I need to check chord shear in a UK Warren truss?

Yes. Chord shear at K gap joints is a critical check that is sometimes overlooked. The vertical components of the brace axial forces produce a shear force in the chord between adjacent braces. For a standard Warren truss with 45-degree brace angles and sinusoidal force distribution (maximum at support, near-zero at mid-span), the chord shear is approximately 70% of the panel vertical shear. If chord shear utilisation exceeds 0.5, the chord face plastification resistance must be reduced by a shear interaction factor per EN 1993-1-8 Clause 7.4.1.2. In deep trusses (span/depth < 10), chord shear is higher because the brace angle is steeper, increasing the vertical component. If chord shear governs the joint design, options include: increasing the chord wall thickness (t0), increasing the chord diameter (d0) to increase shear area, or adding a stiffening plate at the gap location.

What partial factor applies to hollow section joints in UK design?

The UK National Annex to BS EN 1993-1-8 adopts gamma_M5 = 1.00 for hollow section joint resistance. This is notably lower than the gamma_M2 = 1.25 applied to bolted connections. The rationale is that hollow section joints governed by chord face plastification fail through a ductile mechanism (plastic bending of the chord face with visible deformation before ultimate load) rather than brittle fracture. The resistance formulas themselves are derived as lower-bound (5th percentile) curves fitted to the extensive CIDECT test database, providing an inherent safety margin of 1.25-1.50 on the mean test result even before the partial factor is applied. The gamma_M5 = 1.00 combined with the inherent conservatism of the lower-bound formulas produces an overall reliability index of approximately beta = 3.8, meeting the EN 1990 requirement for RC2 structures.

Educational reference only. All design values are per BS EN 1993-1-8:2005 + UK National Annex, BS EN 1993-1-1, and BS EN 10210. Joint resistance formulas are valid only within the parameter ranges specified in EN 1993-1-8 Table 7.1. Designs must be independently verified by a Chartered Structural Engineer registered with the Institution of Structural Engineers (IStructE) or the Institution of Civil Engineers (ICE). Results are PRELIMINARY -- NOT FOR CONSTRUCTION without independent professional verification.

Disclaimer: This content is for educational purposes only. Results must be verified by a licensed professional engineer. Steel Calculator provides preliminary design tools — NOT a substitute for professional engineering judgment.