In 1978 the structural engineer of the Citigroup Center skyscraper learned of a fatal flaw in the design that could cause the tower to topple in high winds. Over the next three months a team raced to secretly repair it at night.
The design of the Citigroup Center (originally known as the Citicorp Center) was an interesting architectural challenge. A church was demolished as part of the construction, but the land was sold to the developers on the condition that a new church would be built on the same street corner where the original once stood. So the skyscraper was designed to arch over this new church, with four columns running not at the corners but the sides of the tower. It looks extremely impressive from the ground level, as if large parts of the lowest floors are missing.
The difficulty with this kind of design is wind. All skyscrapers have to deal with wind to maintain their structural integrity, and this skyscraper accounted for the effects of high winds by directing the forces down into those four columns through a series of “inverted chevrons.” Careful calculations worked out that the building could withstand heavy winds hitting any of the sides of the building. Just one problem: those calculations didn’t account for winds coming in from an angle.
Normally this wouldn’t be a problem, because angled winds carry less force and so less danger than winds hitting straight on. The unique design of the Citigroup Center, those inverted chevrons, meant that the winds coming in at an angle had a much greater effect. Adding to this problem, the joints connecting the chevrons had been downgraded from those in the original blueprints, so they could not withstand the forces now involved. The wrong wind at the wrong time would bring the skyscraper down.
The building’s structural engineer, William LeMessurier, learned of this weakness… after the building was complete and people were working inside. An undergraduate engineering student, Diane Hartley, had made her own calculations and alerted the firm; Le Messurier tested the design and confirmed the flaw.
But LeMessurier did not go public with the immense danger. Instead, building crews spent the next three months retrofitting and reinforcing the skyscraper, welding steel onto the joints so that they could withstand stronger winds. They did their work in secret, at night, to prevent a panic. Very few people even knew that the work was happening, or why it was so necessary.
Fortunately the retrofitting was successful and the risk of collapse mitigated. The extent of the crisis was not known until an exposé by the New Yorker magazine in 1995, and now this whole episode has become a case study in engineering ethics.
[Thanks to Gareth E. for suggesting this topic.]
The Generalist Academy 
When I learned of this years ago, the account (don’t remember where) said that it was actually an engineering student who first discovered the vulnerability and contacted the architect who, to his credit, took it seriously and went to have a closer look, only to discover that the contractor had installed bolts of a lesser grade than the plans had called for.
Gareth E., who suggested this topic, also pointed this out to me this morning. The undergrad engineering student was Diane Hartley, and it may well have been her calculations that set off the whole investigation. This is a longer and more detailed look at the timeline: https://www.theaiatrust.com/whitepapers/ethics/study.php
I’ve updated the article to better account for Hartley’s contribution; I should not have elided that important detail in the first place.