I got this off the Canadian Nuclear Society email list, from one of our members who did her M.Sc. in nuclear engineering at university of Virginia and di much of her course work studying General Electric reactors:
What actually happened was that at one of the plants, the Fukushima Nuclear Plant, when the quake hit, there was a reactor scram and a turbine trip (per design basis) but the emergency diesel did not start. This is a reportable event, it was duly reported, and it resulted in an evacuation.
The plant is a General Electric boiling water reactor, which means that there is no secondary loop - the steam generator is in the top of the pressure vessel, followed by a cyclone separator, and from there the main steam line goes directly to the turbine. The rods come into the core through the bottom, but they are spring loaded and pulled out against the springs, so that on loss of power the poles go into the holes - as they did.
There is a diversion loop going directly to the heatsink for cooling when the plant is not running, and there are also steam driven pumps for circulation - the electric auxiliaries are not critical for heat removal. This is an older plant, with four hours of battery backup available, but note that the steam auxiliaries do not require electrical power to operate - this is deliberate.
At the end of the earthquake the reactor was scrammed, the rods were in, the batteries were on line, the emergency diesel was not. There was several feet (by design, it works this way) of water over the core, giving several hours of boiling cooling before the core would be exposed.
Operating procedure for these plants under such conditions is to spin up the steam pumps, stand by the ECCS (emergency core cooling system) which is nothing more than a pipeline to the nearest available water source, and to get the emergency diesel on line. Note they have four hours to do so. You really don't want to have to use the ECCS, that is not feed grade water, use of it not only gunks up the reactor but results in contamination (the water has to go somewhere) inside the containment.
Here is the conjecture. Unless there was damage to the emergency diesel (minimum, two) this should be doable, so I am assuming it was done. After that, as long as there is sufficient diesel fuel available, things should be under control. Thus, all else is media hysteria.
There's more. Critical loads (such as circ pumps) have more than one source of power. The diesels are designed to start automatically after interruption of switchyard power, and each load (pump) is equipped with an ABT (automatic bus tranfer) device which uses yard power as the primary source but which switches automatically to the secondary source if the primary source is interrupted. Also, everything (remember, this stuff is three phase) is required to be delta connected. That way, if you lose one of the phases you can still get roughly fifty percent (I think the figure is 57 percent) output from the load. Circ pumps are sized with this in mind. That's one.
Then there are the steam pumps. A GE BWR has a torus around the base of the core, filled with steam and water, about 50 percent when things are running well. This can be tapped for steam to run the steam pumps. They can exhaust either to a condenser (if available) or open-loop to the containment. That's two.
Then there is the ECCS. We have discussed this. That's three.
The point is, there is defence in depth in the design. All of the straight mechanical systems can be aligned and operated manually, no electricity required. So cooling should be available.
Having said that, if the diesels truly did come up and then flooded out, they have a serious problem. If they run out of steam (can happen) then the steam auxiliaries will stop and the only thing left will be the ECCS - and that will scrap the core. But that is or should be several days down the road, and if they can get any kind of three phase generator in there (assuming right voltage and frequency), even flying it in, they should be fine.
The ECCS is a flooding system designed to dump cold water directly into the core. If you have to use it then you will crap up the reactor, spill water all over the containment and generally make a mess, but it is just water - any firetruck with a pump can supply enough water to do the job. And if you really get into that much trouble, you could use seawater if you had to. But say goodbye to that core - it will never run again.
But I digress...
