Regarding the on-going crisis of the nuclear plants in Japan, I received this (which I have abridged and reformatted as a numbered list):
So, what do we learn?
Don't use nuclear power.
If you do, don't build in an earthquake zone.
If you do, don't build on a tsunami-prone seaside.
Make sure the emergency cooling system works, even in an emergency.
Make sure the reactor is fail-safe in a power cut.
Of course, if you started with #5, #'s 1-4 would be superfluous.
I think the real lessons look more like this:
Don't be careless with nuclear power.
If you use nuclear power, don't stop your research and development efforts.
If you have to build in an earthquake zone, build to withstand all possible tremblors, landslides and tsunamis.
Make all your plants passively safe (if any are not, see #2).
Nuclear power has saved Japan an immense amount of money for imported fossil fuels, as well as the pollution from burning those fuels and dealing with e.g. ash. Even if several of these plants have to be scrapped, a small radiation release will not contaminate land (as Three Mile Island proved) and will not prevent rebuilding. Japan's nuclear balance sheet probably still shows a fat profit.
The biggest problem may not be the loss of the plants themselves (I understand that they are among the oldest in Japan), but the lack of ability to replace them quickly. The plants which make the large forgings for reactor pressure vessel heads are booked years in advance. If R&D had pushed forward over the last 40 years to yield e.g. small modular molten-salt reactors, it would be much quicker and cheaper to replace the lost plants. There might not even have been an issue, because the oldest, least-robust plants might have been replaced already. But that, sadly, isn't the world in which we find ourselves.
Or for someone to decide that all reactors should have passive safety measures.
It surprises me that there's no such thing as a steam-powered water pump designed for BWRs. It seems rather obvious, and not terribly complex. It could be regulated by a float system and held in the "off" state by applied electric power. Passive safety, voila.
Oddly enough, EP, there *is* - the RCIC or Reactor Core Isolation Cooling system, which is a steamdriven turbo-pump. For some reason they worked at Fukushima Daini 1 through 4 but only at one of the three operational reactors at Daiichi. It's not clear whether they didn't work for some reason or whether they started up but later failed (possibly because they were never run for this long).
The 9.0 Richter earthquake appeared to not compromise even the oldest crappy Fukushima Daiichi reactors. They shutdown just fine and even if you don't shutdown they will just stabilise at a higher temperature level. This is an inherent safety feature of light water reactors - negative power and void coefficients - that makes damage inherently limited.
The diesels started fine and cooldown started. What did the older reactors in was the electrical equipment which flooded after the tsunami hit some time after the quake. So one lesson is to not build critical electrical equipment below grade and make sure the critical equipment is robust and certified water resistant (like submarine electrical equipment certifications).
It is of note that all newer BWRs in Japan are not damaged and worked as intended, and are now in cold shutdown state (<100 degrees C primary water coolant temp).
Basically BWR-5 and up, with the newer Mark II containment and up. All fine.
For example, Fukushima Daiini, the second park of reactors, is also near the epicentre and experienced similar earthquake and tsunami violence. All in cold shutdown mode now, all fine.
So, I'd suspect many of the lessons have already been incorporated in new designs. And BWR-5 isn't really new - 30+ years old...
The newer plants will full passive afterheat removal - no diesel generators to fail, switchgear flooding is no problem - are even better. In fact, the newest design, the Advanced Boiling Water Reactor, is operating at this very moment, in Hamaoka for example, feeding in much needed life-saving electricity.
It seems to me that for newer reactors and for future new build expansion, the issue of learning lessons is only a paper exercise, and no doubt this will be made apparent by the studies which are being conducted at this moment. If modifications have to be made then this must be done to protect the public and convince the public of the low risk of nuclear power.
Ideally, all reactors would be walk-away safe. That's not how things are. The question is, how do we get from A to B?
One necessary condition is for walk-away-safe plants to exist. I believe we've got at least one LWR design certified, and several others demonstrated at the pilot scale (IIRC the EBR-II demonstrated auto-shutdown in a loss-of-cooling test, and the MSRE was even more foolproof).
A sufficient condition for WAS plants to replace the existing fleet is for them to make cheaper power, including the capital cost of new plants. That's much tougher, but MSRs seem to have the best potential for coming close. I've seen projections of $1200/kWe for mass-produced MSRs, which is far lower than anything else out there including ultrasupercritical coal-steam.
"I think a reasonable quid pro quo for the nuke industry is to make all existing nukes walk away safe, or close them down."
Absolutely not. If the generators which would replace the not-quite-as-safe nukes cause more harm to the public, there's both a large financial loss and an externalized loss. The cost to people of economic damage from imported fuels is a factor also. If the alternative to the operational RMBK is coal or imported natural gas, the RMBK may well be preferable for the public welfare. Just don't let idiots mess with it, and replacing the RMBKs should be the highest priority among the nukes! (Replacing coal might be even higher priority.)
"Saying new ones are better is of little help for problems caused by old ones."
Of course not. But the problems prevented by the old ones don't vanish either; if you shut them down without superior replacements, you get net harm.