Hacker News new | ask | show | jobs
by aaron695 1813 days ago
I'm not a fan of the "Hyatt walkway collapse" as a teaching tool. The mathematics is extra interesting. Not sure it's good for the design failure which people reference. More people die from bad insulation design each year, it's the boring stuff that matters.

The Hyatt walkway collapse was at a big dance party with a fair few people watching from above. Perhaps dancing?

I-35 bridge collapse was rush hour.

What's the peak here? Why 1am?

3 comments

> What's the peak here? Why 1am?

I wondered this too. Maybe contraction or rate of contraction?

Sun has gone down, outdoor temperature is dropping, max heat radiation out if the sky is clear, and everybody is home so A/C is on in all units.

The coldest is a bit after dawn. So the previous dawns should have stress tested it. It should fail when the stress is around maximum under this Hyatt peak theory.

Max rate of change of temperature is sunset and sunrise. Rate of contraction I'd expect similar?

I'm not sure AC would change the thermal mass of a building much. One cubic meter of air weights 1.2kg. Changing the concrete walls would be hard.

A drunk hitting a pole at speed in the carpark would fit perhaps. Drunkenness to an empty carpark would be high.

> The coldest is a bit after dawn.

The building (and deck surface) will continue to radiate heat throughout the night on a clear night, especially with few obstructions like a waterfront building has. That's how crops can frost with airtemp > freezing, and why the guidance is to cover your plants. It's not to insulate them against convection.

Counterintuitively in this scenario, this effect is greatest when the wind is calm.

https://en.wikipedia.org/wiki/Radiative_cooling#Architecture

If the peak was actually a trigger, in a residential building 1am is about the start when the max people other than a party will be there (everyone back from clubs)
It's also approximately the time when the building will be coldest. If there were any thermal expansion/contraction issues, they would have been at a maximum near 1am. The thermal expansion and contraction of concrete and reinforcing steel are nearly identical, but the steel has much higher thermal conductivity and cools off far faster; perhaps something snapped under tension between concrete that was still holding heat from the setting sun and steel that had shrunk in the chill of the night.

Incidentally, that chill is also why the 'low battery' alarms on smoke detectors tend to vexingly wake you up in the middle of the night, rather than conveniently running out in the 16 hours a day you're awake, which you might expect to happen twice as often. No, it's not just a bias that causes you to remember waking up to smoke alarm low battery chirps but not hearing them in the daytime, the electrochemical reaction in the batteries happens more readily at higher temperatures. It's a different chemistry, but the same reason a lead acid car battery has a hard time turning a starter in the dead of winter. The sensors in the smoke alarm always draw the same few microamps of current, but when it's coldest in the middle of the night, the battery is less able to meet the demand, and wakes you from your sleep.

Wouldn’t the coldest part of the night typically be immediately before sunrise?
All it takes is being cold enough to cause the extra voltage drop.
The 1978 Citigroup Center engineering crisis is a good teaching tool. The chair of my Uni's Mechanical Engineering department has gotten a good amount of engineering colleges (mainly Mechanical and Civil) to use it as an example/study for students first semester.
What did you learn from the Citigroup Center story?

For me a good story is when a friend built a house the design didn't fit the actual plot, so it was changed by the builders.

One of the final problems was an overhang on the upper roof went over the sky light.

Lessons -

Contrary to common beliefs land plots are not well known.

Engineers/Designers work off incomplete information.

Problems cascade to strange places.