Column-Free Sightlines and Wind-Load Optimization for Stadium Roofs

Span Requirements and Sightline Optimization

A stadium seating bowl is typically 150-250m in diameter for a large venue. The roof must cover this diameter, or a significant portion of it, without interior columns. The span from the perimeter support to the leading edge of a cantilevered canopy is typically 30-40m for a partial roof. For a fully enclosed roof, the clear span across the full bowl diameter can reach 120-200m — a range where conventional double-layer grids become structurally inefficient and beam-string systems or cable-strut systems become the appropriate structural type.

The bolted sphere grid is the workhorse for stadium spans of 60-120m. For a 100m span, a double-layer or triple-layer grid with a depth of 6-8m achieves the span with a steel consumption of 25-40 kg/m². For spans exceeding 120m, the self-weight of a pure grid begins to dominate the structural demand — the grid is carrying its own weight more than the applied loads. At this point, a beam-string system becomes more efficient: the grid forms the upper chord, and prestressed cables replace the bottom chord, putting the grid into compression and the cables into pure tension. This hybrid system, which we deployed for our world-record 229m span coal shed, reduces steel consumption by approximately 30% compared to a conventional grid at the same span.

For cantilevered canopy sections, pipe truss structures provide the linear span and cantilever capacity. A pipe truss cantilever of 30-40m is achievable with a truss depth of 3-4m at the support, tapering to 1-1.5m at the leading edge. The taper is not merely aesthetic — it follows the bending moment diagram, placing material where forces demand it and removing it where they do not. The leading edge of the canopy is the critical wind design zone; vortex shedding at this edge generates the dynamic loads that govern the canopy's structural sizing.

Grid Topology and Spectator Sightlines

The arrangement of grid members — the topology — is not arbitrary. In a stadium, the grid must follow the sightline geometry. If the stadium is oval, the grid follows the oval curvature. If the seating bowl has varying rake angles (steeper in the upper deck, shallower in the lower), the roof's support column positions shift accordingly, and the grid span lengths vary around the perimeter.

A radial grid topology — members radiating from the field center and ring beams at regular intervals — naturally follows the circular geometry of many stadiums. Each radial truss spans from the perimeter to the leading edge, with ring beams tying the radial trusses together and distributing load circumferentially. The bolted sphere system accommodates this radial geometry because each node is independently drilled at the angles the geometry demands — no two nodes in a radial grid share the same bolt angle pattern.

For rectangular or ovoid stadiums, a rectangular grid with varying bay lengths may be more appropriate. The grid is laid out so that column positions coincide with sightline shadow zones, and span lengths vary to accommodate the non-uniform column spacing. The structural analysis must account for the varying spans — a uniform-section grid applied to varying spans results in some members being over-designed (wasting steel) and others under-designed (risking failure). Our optimization process assigns section sizes member-by-member, ensuring each member is sized for its actual force demand.

Wind Tunnel Testing for Cantilevered Roofs

For stadium canopies, code-based wind load calculations are insufficient. The complex geometry of a partial roof — open on one side, solid on the other, with a free leading edge exposed to flow — produces wind pressure distributions that no building code captures accurately. We require wind tunnel testing for all cantilevered stadium roof projects.

A wind tunnel model of the stadium and its surrounding terrain (including adjacent buildings that may channel or block wind) is constructed at 1:200 to 1:400 scale. The model is instrumented with hundreds of pressure taps on the roof surfaces. The tunnel generates wind at varying angles of attack and turbulence intensities, and the pressure taps record the instantaneous pressure at each location on the roof.

The data from wind tunnel testing reveals pressure distributions that are dramatically different from code assumptions. The leading edge of a cantilevered canopy experiences peak suction coefficients of -1.5 to -2.5 (compared to the -0.8 to -1.0 assumed by typical codes). The suction is not uniform — it peaks at the leading edge and decays rapidly within 5-10m. This means the canopy's leading-edge members must be designed for significantly higher uplift forces than the interior members. The structural design uses wind-tunnel-derived pressure coefficients for each grid zone, not a single uniform wind load applied to the entire roof.

Vortex shedding — the periodic shedding of vortices from the canopy's leading edge — generates a dynamic load at a frequency determined by the wind speed and the canopy's leading-edge geometry. If this frequency coincides with the canopy's natural frequency, resonance amplification occurs. The wind tunnel data identifies the critical vortex shedding frequency, and the structural design verifies that the canopy's natural frequency is sufficiently removed from it.