> He claims that TMI showed that there is no risk of nuclear fuel melting through the steel containment structure. But at TMI, the core was actively cooled through the entire incident, and the temperatures were not even high enough to melt the fuel itself.
Actually, about 1/3 to 1/2 of the Three Mile Island core melted completely (i.e. far worse than mere zirconium cladding damage).
This was because the Three Mile Island core was not continuously cooled. There were a couple of elements to this:
1. In normal operation, the primary system is kept highly pressurized to ensure that the primary coolant does not boil. The way this is done, however, is to use something called a pressurizer, which does have a steam bubble, for much the same reason cars use shock absorbers. When the steam flow was cut off to the turbine generators, that acted as a loss of the heat sink that had been cooling the reactor, which caused pressure to go up.
When pressure got too high, a relief valve in the pressurizer automatically lifted to relieve the pressure increase and maintain primary piping integrity. The valve was supposed to automatically shut once the pressure dropped enough, but did not. The indication that the operators had showed that the valve had been commanded to shut, but did not show that the valve had not actually shut. The location of the leak tended to artificially increase water level in the pressurizer by essentially "sucking" it from the reactor vessel (the operators had no indication of reactor vessel water level, since it was normally completely full of liquid).
This loss of water from the reactor meant that it uncovered the fuel much quicker than might otherwise be expected. The operators had other ways of telling something was wrong (such as the sharp pressure drop from a stuck open relief valve) but did not put the puzzle together. They even did worse: Automatic safety systems kicked on to force-fill the reactor with cooling water, but the operators were afraid the system would overflow the entire primary (remember, they thought there was still water in the reactor vessel), so they dampened the flow, then shut off the system completely.
2. As steam started forming in the reactor vessel, the coolant loop piping connecting the reactor vessel to the reactor cooling pumps started to fill with steam, which caused the reactor cooling pumps to cavitate (which is certainly not good for the pumps). The operators had indication of this, but assumed the cavitation was due to the pressure drop and, again thinking the core was still covered with water, they shut down the reactor cooling pumps. This represented the loss of the last line of defense against reactor meltdown.
It was not until some hours later when a different shift relieved that it was realized that a relief valve was stuck open and that there was no water over the fuel. They started re-established cooling at this point, but the damage was already done.
Ok, I misremembered and my comment was wrong. For about 40 minutes all the cooling pumps were turned off, and during the following 7 hours cooling was ineffective due to hydrogen bubbles. (http://www.threemileisland.org/downloads//195.pdf). The core partially melted. Sorry.
This doesn't make the original article any more correct though. During most of the TMI accident the core was being actively cooled, and in fact it never got uncovered enough to even completely melt. So this still does not tell us anything about how containment would work in the absence of cooling.
> During most of the TMI accident the core was being actively cooled.
Steam is not a coolant, which is why I went into so much detail on my first point. In fact steam has such atrocious heat conduction properties compared to water that you'll find many nuclear engineers refer to "steam blankets".
So if the coolant pumps were running or not is essentially immaterial if the coolant flowing over the fuel plates is itself steam. I only said the coolant pumps were the last line of defense. They are required, but not sufficient.
Actually, about 1/3 to 1/2 of the Three Mile Island core melted completely (i.e. far worse than mere zirconium cladding damage).
This was because the Three Mile Island core was not continuously cooled. There were a couple of elements to this:
1. In normal operation, the primary system is kept highly pressurized to ensure that the primary coolant does not boil. The way this is done, however, is to use something called a pressurizer, which does have a steam bubble, for much the same reason cars use shock absorbers. When the steam flow was cut off to the turbine generators, that acted as a loss of the heat sink that had been cooling the reactor, which caused pressure to go up.
When pressure got too high, a relief valve in the pressurizer automatically lifted to relieve the pressure increase and maintain primary piping integrity. The valve was supposed to automatically shut once the pressure dropped enough, but did not. The indication that the operators had showed that the valve had been commanded to shut, but did not show that the valve had not actually shut. The location of the leak tended to artificially increase water level in the pressurizer by essentially "sucking" it from the reactor vessel (the operators had no indication of reactor vessel water level, since it was normally completely full of liquid).
This loss of water from the reactor meant that it uncovered the fuel much quicker than might otherwise be expected. The operators had other ways of telling something was wrong (such as the sharp pressure drop from a stuck open relief valve) but did not put the puzzle together. They even did worse: Automatic safety systems kicked on to force-fill the reactor with cooling water, but the operators were afraid the system would overflow the entire primary (remember, they thought there was still water in the reactor vessel), so they dampened the flow, then shut off the system completely.
2. As steam started forming in the reactor vessel, the coolant loop piping connecting the reactor vessel to the reactor cooling pumps started to fill with steam, which caused the reactor cooling pumps to cavitate (which is certainly not good for the pumps). The operators had indication of this, but assumed the cavitation was due to the pressure drop and, again thinking the core was still covered with water, they shut down the reactor cooling pumps. This represented the loss of the last line of defense against reactor meltdown.
It was not until some hours later when a different shift relieved that it was realized that a relief valve was stuck open and that there was no water over the fuel. They started re-established cooling at this point, but the damage was already done.