Comment by aftbit | Hacker News Reader

Comment by aftbit | Hacker News Reader

Parent Full thread

>Because each and every incident can be traced back to ignoring whatever safety standards and practices where applicable at the time these accidents happened.

No they absolutely cannot. Many incidents happened due to yet unimagined failures. For example, KLM Flight 867, which lost all four engines due to a common mode failure (flying through ash clouds). The plane landed safely and safety standards were updated to provide guidance on avoiding, detecting, and reacting to ash ingestion.

TMI was a similar situation. At that point, nobody had really considered human factors engineering for nuclear plants. Nobody considered the risk of having a lamp that indicated that a valve had been asked to close instead of that it actually was closed.

One major challenge that aviation has handled better than nuclear and much better than chemical production is ensuring that issues get reported promptly and treated seriously. Aviation calls it "just culture", where mistakes and accidents are treated with retraining and lenience, but covering up issues is treated very harshly. This is absolutely necessary, otherwise you end up with major communications issues like we saw with MetEd during TMI or Tepco during Fukishima.

I ment nuclear accidents and disasters, both Chernobyl and Fukushima can be traced back to ignoring safety rules and regulations.

Ah I see. Let me write up a few thoughts about the "big 3" nuclear accidents.

Chernobyl was a very poor design, with a positive void coefficient of reactivity (i.e. temperature goes up, steam goes up, reactivity goes up, leading to runaway feedback) and no containment building. It was operated in a very poor manner as well, which was the principle cause of the failure. Some sources[1] claim that this poor design and unsafe operating conditions were well known to plant management and Soviet leadership, and were covered up in the name of Progress and production pressures. There was even a very similar accident that occurred during a very similar test in Leningrad in 1975, albeit with less extreme consequences, which was fully covered up, even from plant operators at other RBMK reactors.[2]

Fukushima was a pretty well designed and operated reactor, but the design basis did not consider the risk of a large tsunami, rather instead focusing on typhoons. Ironically, they actually removed a larger "natural sea wall" (i.e. a cliff) that would have protected the plant.[3] The big lesson learned here is to have prepositioned stocks of generators, batteries, replacement parts, etc close enough to the plant to be supplied promptly as needed, but far enough away to hopefully be excluded from any local disasters, aka the FLEX program.[4]

On the other hand, TMI was an accident in which the operators made the wrong call, primarily because they had been trained to be concerned about one specific danger ("going solid" and bursting the pressurizer, requiring less water injection), but really they were facing a totally different one (small break loss of cooling, requiring more high pressure water injection).[5] Interestingly, a very similar accident happened at Davis Besse in Ohio only a few years before, but without any major consequences, as the operations team recognized the mistake and resolved it, but these lessons learned were not well communicated to operators of other plants. This failure and others like it led to the establishment of WANO (international) and INPO (USA) which are organizations intended to help operators share experiences with each other in a timely and safe manner.[6][7]

1: https://www.reuters.com/world/unsealed-soviet-archives-revea...

2: https://en.wikipedia.org/wiki/Leningrad_Nuclear_Power_Plant

3: https://oilprice.com/Latest-Energy-News/World-News/Tepco-Rem...

4: https://www.nrc.gov/docs/ML1524/ML15244B006.pdf

5: https://www.ans.org/news/article-1556/tmi-operators-did-what...

6: https://www.wano.info/services/overview-of-services-(1)

7: https://www.inpo.info/history

And the estimated cleanup costs are ~500 billion, ~50 billion and ~1 billion. Expensive mistakes.

The cleanup costs are high enough that you could have literally replaced the entire nuclear power plant fleet of Japan and built new ones. When it comes to nuclear safety, tear it up and build again is cheaper than any single accident. Nuclear power plant advocates don't seem to understand the magnitude of the failure of a nuclear power plant. It simply is unacceptable.

Two additions to your otherwise great summary:

After the Leningrad incident, the RBMK Chief Engineering organisation announced changes to the operating procedures taken this incident into account to all RMBK plants. Those procedures where never written, and none of the operators asked about them neither. That was pure negligence, and no cover up (the cover up was invented by the HBO series, no idea why). What is true so, is that the whole bureaucracy and organisation in the USSR let Chenobyl 4 get away with well known safety and regulation violations before the accident, and those regulation were extremely lax by modern standards to begin with.

An interesting bit about Fukushima is not the reactor design, there is apparently nothing inherently wrong with that, but rather the site. As you said, the seawall was too low, mainly because during the site development earthquakes of the magnitude that caused the tsunami where not considered despite geological proof to the contrary. That was negligence number one. Number two occured after simulations years before the tsunami predicted pretty exactly the magnitude of the earthquake and the height of the following tsunami. Despite being fully aware of that simulation, and the effects on the site, no measures where taken, e.g. updated procedures or a heigher sea wall or moving batteries and generators to higher ground.

The IAEA reports on both, Chernobyl (the IAEA report in English includes translations of the two Soviet ones, and man are those two damning even before you consider who wrote them) and Fukushima are very good reads, highly recommended if you didn't read them already.

Reply on news.ycombinator.com