[falcolas]

> improve the structure

The key question is: Who would have paid for it?

Once running, much of the structure is considered to be radioactive. Work in active radioactive areas is expensive and time consuming. Without shutting down the reactor entirely followed by an expensive cleanup, doing major structure upgrades to the reactor chambers is virtually impossible.

They also couldn't just decommission it once they realized their errors and rebuilt another one somewhere else - again the costs would have been astronomical.

People are terrible at managing (what was then) theoretical risks. If you can't convince people and companies to move out of proven high-risk earthquake areas, imagine trying to convince tax payers to pay billions to upgrade or move a functional-but-theoretically-vulnerable power station.

[WalterBright]

> Who would have paid for it?

The disaster could have been prevented by the simple expedient of elevating the backup generators, or putting them in flood-resistant bunkers. Cheap.

[inimino]

Hindsight.

The actual tradeoff would have been not just this one mitigation, but all the mitigations for all the equally probable risks. Including all the assessments to find out what those risks would have been.

[discoursism]

I don't think you can say, "hindsight," when the mitigation was actually recommended in reports that were made before the event. It's only hindsight when mitigations become obvious after a disaster.

[krallja]

The hindsight is in which mitigation would have been effective. In a different disaster, perhaps the generators would not have been the missing piece.

[discoursism]

It's only really possible to make a determination about whether it was hindsight or not after seeing the prominence of the recommendation. Suppose I tell someone a thousand times, "you should wear seatbelts," and once to fix their taillight. Then they are killed in an accident after being flung from a vehicle while not wearing a seatbelt. It would not be hindsight for me to say, "they probably should have worn a seatbelt." Maybe it would be if they were killed in a freak accident involving an unfixed taillight.

Unfortunately, given I have only second-hand knowledge of these reports' existence, and given they are probably not in my native tongue, I'm not really in a position to validate what the situation was vis a vis the generators. However, generators positioned on the coast, behind a low seawall, in a country that is regularly inundated by tsunamis, appears to be an entirely predictable failure mode. Not quite to the degree of not bringing a parachute on a skydive, but in that direction.

[WalterBright]

I agree, I think that the language barrier is not a problem when you see those generators sitting right behind the seawall.

It's like in New Orleans, where the emergency generators were located in the basements below sea level. There's nothing subtle about that. Sorry, but how daft can one get?

[WalterBright]

> Hindsight

I disagree. Failure analysis does not start with "assume the seawall will never fail." It starts with "assume the seawall fails. What are the consequences?"

Given the location of those generators, it was obvious that a breach would take out CRITICAL generators, and it would have been cheap to remove that vulnerability.

Another cheap remedy would have moved the hydrogen vent pipe to vent outside, rather than INSIDE AN ENCLOSED SPACE FILLED WITH SPARKING ELECTRICAL EQUIPMENT.