Pretty sure it's actually extremely interesting.
The test was considered such a non-risk that it required next to no oversight [1] [2]. If something that doesn't require oversight results in a nuclear disaster then something is wildly wrong with your regulations and design.
[1]: https://en.wikipedia.org/wiki/Chernobyl_disaster#Safety_test
[2]: https://www-pub.iaea.org/MTCD/publications/PDF/Pub913e_web.p...
You missed the from an engineering perspective part at the begining of this sentence.
Perhaps _now_ it's less interesting as the faulty design has been studied at least twice with computer simulations (and a lot more skepticism on the initial Russian presentation which blamed the operators).
[1]:https://www.iaea.org/publications/3598/summary-report-on-the...
[2]: https://www.iaea.org/publications/3786/the-chernobyl-acciden...
TMI also wasn't mainly an engineering problem: it was the way the incident reaction guidelines where written and how crisis group was set up that was the largest cause, and as such there has been little engineering changes following TMI but a lot of organizational changes.
But it has been addressed with governance rules, in my country's power plant the guy responsible for safety is hierarchically independent from the guy responsible for running the plant so the later cannot command the former and let his hubris destroy the plant.
With more heat the boron atoms (and water) get further apart. Fun times.
> the high cost mostly comes out of the fact that we don't build nuclear reactors as series + the fact that we finance it at insane rates
The UK has had a hard time getting this. Sizewell C is still pending a final investment decision (while we risk the loss of experience from Hinkley Point C), although the Regulated Asset Base model may help reduce the finance rates.
China appear to get it, mind you. 11 new reactors forecast to cost 33B USD, using experienced teams. But they have energy security issues.
Not only this, but with the Doppler effect, their ability to capture neutrons diminishes as well.
Positive feedback loop isn't needed for a nuclear accident to happen. Sure it's what happened in Chornobyl, but not in TMI or Fukushima. And from an engineering perspective Chornobyl isn't that interesting as an accident example because it's mostly a product of brainwashed egotic manager who had all the power over the engineers.
Also it's not always entirely straightforward to keep the void coefficient negative at every point of the operating cycle, especially if things go wrong: PWR have a negative void coefficient most of the time but not 100% of the time: when the reactor is cold you put tons of boric acid into the water to counteract the reactivity and avoid divergence, but at this particular time the void coefficient is positive because of the high level of Boron. Of course in regular events it doesn't matter because the reactor is off, but that's something that can also happen during an emergency situation where you inject a massive amount of boron in the water (there are scenarios where you do that).
But again, the reactor's power getting out of control isn't the biggest risk anyway, the biggest problem comes from the fact that residual power is still annoyingly high even when you've shut down your reactor and you need to deal with it. The fact that you can't just shut it down and everything's OK when something is wrong is the real pain of working with a nuclear reactor.
Source: I have a nuclear engineer specialized in immediate response to incidents and accidents at home.
And the high cost mostly comes out of the fact that we don't build nuclear reactors as series + the fact that we finance it at insane rates. Antinuclear activists have their responsibilities in that, but even without them I suspect most states wouldn't be doing the right thing either: nuclear isn't a good fit for neoliberal thinking anyway.