IDA bottling plant — a structural reassessment

Context

The former IDA mineral-water bottling plant — a large-span prefabricated concrete hall from the 1970s — was bought at auction to serve as a warehouse. Before reconstruction, a highly respected local senior engineer assessed the structure and found it in a critical, near-failure condition: the roof, he concluded, was carried by post-tensioned prefabricated trusses of the kind that had collapsed without warning elsewhere in the country. His remedy approached the cost of a new hall — hundreds of thousands of euros, potentially over a million.

I was brought in to develop a remediation strategy alongside him.

A warning with a history

The alarm he was responding to was real. In 2010 and again in August 2018, a roof truss collapsed in the same industrial complex in western Bohemia — both type SPP 6-18/6 from the ZIPP prefabrication system, 18-metre girders assembled on site from three 6-metre segments and clamped together by tendons stressed during erection. In both cases the tendons had corroded inside ungrouted ducts and the truss failed by brittle fracture. ČKAIT, the Czech Chamber of Chartered Engineers, held a press conference in December 2018, issued a call to its members in 2019 and passed the matter to the Ministry of Regional Development, which had building authorities notify the owners of affected halls. More than half a million square metres of roof carried by this truss family were documented, with an estimated 13,000 to 20,000 people working underneath.

So a senior engineer who looked at a 1970s prefabricated hall with concrete lattice trusses and thought of SPP 6-18/6 was not being careless. The trouble is that the warning concerns a specific product — and the 1970s are exactly the decade in which that product was replaced.

Two technologies, one era

The segmented ZIPP trusses were approved for series production in 1962. Around 1975 the 18-metre type went out of production, superseded by trusses cast in one piece and pre-tensioned in the factory. A hall built in the 1970s can carry either kind, and to a casual glance they look alike. Mechanically they could hardly be more different.

In a post-tensioned segmented truss the segments are cast with empty ducts through both chords and the end diagonals. On site, bundles of 4.5 mm wire are threaded through, stressed against anchor plates at the ends of the girder and — in theory — grouted. That steel, about 2% of the truss's mass, carries essentially all of its tension.

Where grout is missing, the wire sits in a void that collects condensation and roof leaks, right at the upper anchorages under the gutter units. Corrosion at one point releases the tendon along its whole length; there is no ductile reserve and nothing to crack visibly first. Surveys have found grout absent or defective in most ducts inspected, and no non-destructive method can certify the wires sound.

In a pre-tensioned truss the wires are stressed on a casting bed before the concrete is poured and bond to it along their full length. No ducts, no voids, no anchor plates, no grout to forget. Local corrosion weakens a wire locally rather than releasing it, and the concrete protects the steel as it does ordinary reinforcement. Durability becomes the ordinary question for any concrete element — cover, carbonation, cracking — inspectable by conventional means.

The infamous failure mode, in other words, belongs to one construction method and, in this country, essentially to one product family. Everything depended on which technology the IDA hall actually used.

Reading the trusses

Rather than accept the diagnosis, I set out to establish the product these trusses were, not the family they resembled. The segmented type has a recognisable signature — verticals in the web, doubled where segments meet every six metres, a joint line across both chords at the same points, anchor plates at the ends of the bottom chord.

Seen from the floor, the verticals are the first thing that registers.

The IDA trusses have none of it: a diagonal-only web, chords continuous from support to support, clean concrete at the ends.

A visual reading is not proof — repainting can hide joints, and the literature warns that misidentification is common. So I went to the documents: the original structural drawings, the period type catalogues and the surviving records of the plant that produced the elements. They confirmed what the geometry suggested. The IDA trusses are type SPP 10-18/6: one-piece girders, pre-tensioned in the factory with wires anchored by bond. The failure mode the alarm was built on could not occur in them.

Verification and outcome

A conclusion that overturns a senior colleague's verdict should not rest on one person's reading of an archive. I commissioned an independent inspection and diagnostic survey from the specialist firm whose engineers had developed the national methodology for identifying and securing post-tensioned trusses — nobody was better placed to disagree with me. Their survey confirmed the identification in full. Presented with the evidence, the original consulting engineer reviewed it and openly acknowledged the correction.

The reconstruction proceeded as a standard renovation of an old but sound hall, the intervention originally recommended — effectively a new roof structure — no longer on the table. As an unexpected personal outcome, the firm that verified my analysis subsequently offered me a position.

Reflection

Engineering authority is real and deserving of respect — and fallible, most of all when it is applying a genuine warning. The ČKAIT alarm was correct. The error was one of resolution: a diagnosis by family resemblance is a different thing from a diagnosis of the exact product in front of you, and the two diverged in precisely the decade this hall was built. Questioning a senior verdict does not require arrogance; it requires the patience to read the catalogues, find the production records, and then have the result checked by the people best qualified to disprove it.