Guodian Ningxia Fangjiazhuang Power Plant Enclosed Coal Shed

No client asks for a 229-metre span out of ambition. They ask because the alternative — a narrower shed with interior columns — fails on four hard grounds.

1. The stockpile and the machinery that must cover it

A large thermal power unit burns on the order of 1.5–2 million tonnes of coal per year, and a storage yard sized for meaningful strategic reserve and procurement flexibility is measured in hundreds of thousands of tonnes. That inventory forms a stockpile whose width and height are dictated by conveyor geometry and reclaim strategy. The reclaim system is typically a stacker-reclaimer on rails — a machine with a boom reach of tens of metres that must sweep the entire footprint of the pile to deliver fuel to the plant bunkers. Every square metre the machine cannot reach is coal that must be reclaimed by smaller, slower equipment or by dozers zoned around the rails. Columns inside the envelope would riddle the reclaim machine’s travel corridor and blind spots into its path — an operational impossibility at a plant whose fuel supply cannot be interrupted.

2. Environmental regulation changed the rules

Air-quality compliance under Chinese environmental law requires thermal power plants to store bulk coal in enclosed facilities — wind-borne coal dust from open yards is a measurable fugitive-emission problem, and regulatory enforcement has made open storage progressively untenable. Enclosure must cover the stockpile, the reclaim rails, the transfer towers and the yard-crane envelope. Once the enclosure must span the reclaim machine’s full travel width plus the pile’s clearance zones, the building envelope is already at the scale of a very large span — and the cost equation flips: beyond roughly 60–80 m of clear width, a grid roof or arch system with no interior columns becomes cheaper and faster to build than a forest of internal columns plus their local enclosures, foundations and dust-control headaches.

3. Stacking height drives height; span drives economy

The shed height is set by the reclaim boom and the pile’s natural repose angle — typically 35–40 degrees for thermal coal. The width is set by inventory and machine travel. Both are dictated by physics and operations, not by architecture. Given a target clear interior width of roughly 220 m for the yard envelope, the shed either spans it with one structural system or with a sequence of narrower vaults interrupted by internal supports. The latter multiplies foundation, membrane, drainage and maintenance lines and re-introduces the machine-obstruction problem. A single continous span is not luxury; it is the low-total-cost solution.

4. The record was a boundary of capability, not a marketing target

Engineers on both sides of the table understood that crossing 229 m with a passive arch or truss would have required either absurdly heavy members or unacceptable deflection. The beam-string system with prestressing existed precisely to break that deadlock — it had already demonstrated viability at spans approaching 200 m elsewhere, and the project team’s job was to take the system into the 220 m+ class with complete analytical honesty. The world-record status was a by-product of solving the client’s problem correctly; it was not the design objective. That distinction matters, because it is the difference between a structure built for a photograph and a structure built to earn coal for twenty years.

Prestress is a stored load. Getting it right — and proving it on site — is the heart of a beam-string project, and the discipline is different in kind from ordinary steel design.

Analysis that takes the cable seriously

The structure was analysed with geometric and material nonlinearity from the first day, because cable action is inherently nonlinear: cable sag, cable elongation, strut shortening, rib shortening and temperature all interact with the prestress level, and the deformed geometry of the final structure differs measurably from the un-tensioned geometry. The working basis was a full three-dimensional finite-element model in which every rib, strut, cable and brace participated, with:

• Two-stage analysis: Stage 1 — the structure under dead load plus the prescribed prestress force, establishing the deformed geometry and member forces at “state of equilibrium after tensioning”; Stage 2 — all subsequent load cases (live, coal-pile asymmetric load, wind, snow, temperature, seismic) applied to the pre-stressed state.
• Asymmetric live load: the deadliest case for a long-span arch is not uniform load but the partial load of an asymmetric stockpile — coal piled high on one side, empty on the other — combined with wind. Every intermediate filling condition was enveloped, because the shed spends its life in partial fill, not in the neat full-symmetry state textbooks draw.
• Stability with geometric imperfection: global and member-level stability were verified against the code imperfection envelopes for grid structures of this class, with the prestress state included as part of the stiffness reference.
• Temperature: a shed of this scale experiences large thermal swings between winter and summer and, in Ningxia, between day and night; cable force and rib stress were enveloped across the seasonal and diurnal temperature range, and movement provisions were placed accordingly.

The tensioning sequence

Prestress is never simply “applied”; it is staged and verified. On this project the cable systems were tensioned in a defined sequence from the abutments toward the crown zone, strand by strand in the multi-strand anchorages, using calibrated jacks with simultaneous control of jack force and strand elongation. The acceptance criteria were dual — force and elongation must both land inside their envelopes — because elongated-but-undertensioned strands and tensioned-but-underelongated strands are each failures of a different kind. After tensioning, cable force was further verified by the frequency (vibration) method on sampled strands: the measured natural frequency of the free cable length yields the tension directly, an independent cross-check on the jack readings.

The sequence was staged so that each rib reached its completed prestress state while the adjacent ribs and the longitudinal bracing could absorb the redundant force redistributions — on a beam-string shed, tensioning one rib changes its neighbours’ state slightly, and the model predicted exactly how much. The survey verification of the crown line after tensioning matched the predicted deformed geometry within the specified tolerance, which is the moment a beam-string structure proves itself: the steel does what the model says, the crown rises by the computed value, and the “record span” becomes a measured, documented fact rather than a headline.

A passive 229 m arch pushes outward on its supports with a horizontal thrust measured in thousands of tonnes. That thrust must be resisted — by massive abutments, by endless ties, by pile groups designed for uplift and drag — and the cost of resisting it invisible in the roof photograph. The beam-string system closes the force loop instead: the tensioned cable runs between the two abutments, so the horizontal component of the arch’s action is carried inside the structure itself, and the foundation experience is predominantly vertical load.

That property was not decorative on this site. The shed sits within a working power-plant yard where the ground under and around it carries buried services, drains, conveyor tunnels and hard-standing; a structure that demanded broad structural abutments and deep tie systems would have collided with everything the yard needed to keep doing. With the self-balanced system, the foundation scheme was concentrated, predictable and cooperative with the site’s existing infrastructure — a design benefit expressed in metres of dug earth, tonnes of concrete and weeks of programme.

The abutment details themselves carried the discipline inward: the cable anchorage zones were designed as bearings tolerant of the multiple force directions the prestress programme applies, verified by NDT, and detailed so that the anchorages remain inspectable for the life of the shed. Every hot-spot at the ends of a self-balanced system is hidden in plain sight — deliberately designed, deliberately inspected, deliberately documented.

No crane in the world lifts a 229 m arch as a single piece, so the structure was assembled from the ground up, in a sequence that kept every intermediate state stable and verifiable.

Ground assembly and block lifting

Rib segments were assembled at ground level on purpose-built jigs — where welders work at bench height, inspection is accessible, and fit-up tolerances are checked against the surveyed jig before any permanent welding. Completed segments were then lifted by heavy crawler cranes onto temporary support towers placed along the span, docked, aligned and connected. Temporary towers were staged so that each new segment’s weight, plus the wind load on the partially erected ribs, was fully accounted in the erection scheme — with temporary guys and plan bracing detailed in the drawings, not improvised in the field.

The closure segment

The final segment — the crown closure — was dimensionally verified against the two sides’ accumulated geometry before cutting. Closure on a 229 m arch is the moment where every fabrication tolerance, every lift settlement and every thermal displacement of the day expresses itself in a single joint; the procedure therefore specified the acceptable gap window, the time-of-day targeting for temperature stability, and the sequence of stitch-welding to closure. It was executed once, correctly, which is what happens when the closure geometry is analysed as carefully as the service load cases.

Tensioning as the final erection step

After the full rib set was self-stable, the prestress programme ran in the staged sequence described above — abutment-to-crown across the rib population, with dual force-and-elongation control. The final evidence of success was not a jack gauge but the survey: the crown line settling onto the design surface, the cable forces confirming the model, and a 229 m structure standing with its own internal tension balance — no external tie-downs, no visible props, no residual wherewithal. The enclosure and cladding followed, converting the skeleton into the dust-tight, weather-tight storage envelope the plant operates today.

A record span earns credibility through measurement, and this project was instrumented and verified on three levels.

• Construction-stage verification: every tensioning stage was documented with jack force and elongation; sampled cable strands were verified by frequency-method tension measurement; and crown-line survey after tensioning confirmed the predicted deformed geometry.
• Commissioning and early-load verification: the shed was operated through partial-fill states while the structure was surveyed again — the asymmetric-stockpile case that dominated the analysis was examined live, against model prediction, and found to match within the specified margins.
• Continued service record: the enclosure has since served its power-generation duty through multiple annual cycles — hot summers, the severe cold and wind of Ningxia winters, repeated asymmetric filling and reclaiming — without structural incident, with the cable system exhibiting stable tension and the connections free of distress at inspection.

That record is the reason the 229 m span is presented in this portfolio as engineering evidence rather than as a trophy: it was built to be measured, and it measures true.