The 229m Engineering Frontier: Why Coal Sheds Need Ever-Larger Spans

The Structural Efficiency Threshold

A bolted sphere spatial grid structure is the workhorse for coal shed spans up to approximately 150m. At these spans, a double-layer or triple-layer grid with a depth of 8-12m achieves the clear span with a steel consumption of 30-50 kg/m². The grid's self-weight is a manageable fraction of the total load, and the member forces are within the range of standard steel sections.

Above 150m, the grid's self-weight begins to dominate the structural demand. The top chord — carrying compression from the grid's bending under gravity — must resist forces that increase with the square of the span. At 200m, the top chord compression may be 2.5-3 times that of a 120m span grid, requiring significantly heavier sections. The bottom chord, carrying tension, faces a similar proportional increase. The grid's depth must increase to 12-15m to maintain the depth-to-span ratio, further increasing self-weight. The result is a structure where 60-70% of the steel is carrying the grid's own weight, not the applied loads — a diminishing return that makes pure grids structurally and economically inefficient at 200m+.

The Beam-String Solution

A beam-string structure solves this problem by replacing the grid's bottom chord with prestressed high-strength steel cables. The cables, tensioned to a predetermined force, put the upper grid into compression and themselves into pure tension — the ideal stress state for each material. Steel excels in compression when braced against buckling (which the grid's web members provide). Cable excels in tension, with no buckling concern and a strength-to-weight ratio 3-5 times that of structural steel.

The result: at 229m, the beam-string system consumes approximately 30% less steel than a conventional grid of equivalent span. The cables replace hundreds of tonnes of bottom-chord steel with a few tonnes of high-strength wire. The upper grid, now in compression, is braced by the web members and the cable's tension, preventing buckling. The structural depth is maintained at 12-15m, but the mass within that depth is dramatically reduced.

The Pre-Tensioning Sequence

The beam-string system's structural integrity depends on the cable pre-tensioning. The cables must be tensioned to a force that achieves the design compression in the upper grid under permanent dead load. Too little tension, and the upper grid goes into tension under live load — a state it was not designed for. Too much tension, and the grid is over-compressed, risking buckling.

The pre-tensioning sequence for our 229m coal shed was engineered in stages:

• Stage 1 — Grid assembly: The upper grid and web are assembled on the ground or on temporary shoring, with the cables installed but slack. The grid geometry matches the design pre-tension shape — not the final dead-load shape.
• Stage 2 — Initial tensioning: Cables are tensioned to approximately 30% of design force. This lifts the grid off the shoring and puts the web members into a stable stress state. The grid geometry is surveyed and compared to the model — deviations indicate fabrication errors or cable length discrepancies that must be corrected before further tensioning.
• Stage 3 — Intermediate tensioning: Cables are tensioned to 60-70% of design force in a synchronized sequence — all cables at a given cross-section are tensioned simultaneously to avoid generating unbalanced lateral forces on the grid. The synchronization is controlled by computer-monitored hydraulic jacks that measure both force and elongation at each cable anchor.
• Stage 4 — Final tensioning: Cables are tensioned to 100% of design force. The grid deflects to its final position under the cable tension. The total deflection is compared to the predicted value from the structural model — agreement within 5% confirms that the as-built structure matches the as-designed model.
• Stage 5 — Cladding and dead load application: Roof cladding, wall panels, and permanent dead loads are applied. These loads increase the grid's compression and the cables' tension — both within the design envelope. Final cable tension is verified by vibration-based force measurement, which infers cable force from the cable's natural frequency.

Throughout the sequence, cable tension is monitored by load cells at the anchor points and by vibration-based force measurement. Any deviation from the predicted tension-elongation relationship triggers a halt and engineering review. The pre-tensioning sequence is not a construction operation — it is a structural engineering process that determines whether the finished structure will perform as designed.

A 229m clear-span coal shed is, from a wind engineering perspective, a very large enclosed shell with one or more large openings for truck and train access. The wind does not treat it like a typical building — the long span, the large interior volume, and the partially open end walls create pressure distributions that building codes do not accurately capture.

For the 229m shed, we conducted wind tunnel testing at a scale of 1:200. The testing revealed two critical findings that code calculations missed:

First, internal pressure amplification. When wind blows against the open end of the shed (the end with the truck/train access door), the building acts as a pressure vessel: wind enters through the opening and pressurizes the interior. The internal pressure acts outward on the roof and walls, combining with external suction to create a total uplift force significantly higher than the code-predicted value. The design uplift at the roof's leading edge was 1.8 times the code value — enough to change the structural sizing of the leading-edge members.

Second, vortex shedding on the long roof. A 229m-long roof with a curved or gabled cross-section sheds vortices at a frequency determined by the wind speed and the roof's crosswind dimension. These vortices generate alternating pressure zones on the roof surface, creating a dynamic load that can excite the roof's lateral vibration modes. The wind tunnel data identified the critical vortex shedding wind speed and the resulting dynamic amplification factor, which was incorporated into the structural analysis as a dynamic load case.

The wind load design also addressed the partially built state. During construction, the shed may have cladding on one end and the other end open — a configuration that generates internal pressure forces very different from the completed building. The construction-period wind load case was checked separately, and temporary bracing was specified for the open-end condition to prevent damage during the cladding installation sequence.