The 8.9 magnitude earthquake in Japan is causing problems for at least one of its fleet of nuclear reactors—and authorities have shut down 10 of the country's 55 units. Tokyo Electric Power confirmed that pressure had been rising inside reactor No. 1 at its Fukushima Daiichi nuclear power plant on the northeast coast, one of the largest nuclear power plants in the world. That means cooling water is not getting to the reactor core, causing a build up of steam inside the containment vessel. The problem, according to Japanese media reports, is a loss of grid electricity to run the pumps that bring in cooling water. The backup diesel generators that are supposed to provide emergency power in that case are out of order, according to the Japan Atomic Industrial Forum, but replacements were being taken to the plant. (Similar diesel generators were providing power to the nation's Rokkasho Reprocessing Plant, which recycles spent nuclear fuel.)
As a precautionary measure, the Japanese government has declared a nuclear emergency and asked people living within three kilometers of the facility to evacuate and people living within 10 kilometers to remain indoors. Tokyo Electric Power, for its part, planned to vent some of the radioactive steam from inside the containment building.
Scientific American spoke with Scott Burnell, public affairs officer at the U.S. Nuclear Regulatory Commission (NRC), the government agency charged with monitoring the safety of the 104 nuclear reactors in the U.S., about what it takes to cool down a reactor.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
[An edited transcript of the interview follows.]
How do you typically cool a reactor?
The approach to cooling is very simple: push water past the nuclear core and carry the heat somewhere else. The chain reaction that actually runs the reactor can be shut off in a matter of seconds. What's left over in the core, the radioactive material, will continue to give off heat for a long time. Unless you have a mechanism to remove that, the heat can build up and can eventually damage the radioactive fuel or the reactor.
Pushing water past the core means pumps that are generally run by electricity. What happens when a reactor gets disconnected from the grid?
There are emergency diesel generators. You also have a battery system to keep instruments running, but that can also provide power to safety systems [which prevent a meltdown by cooling the reactor core]. It's all meant to provide defense in depth. First you rely on the grid. If the grid is no longer available, you use diesel generators. If there is an issue with the diesels, you have a battery backup. And the batteries usually last long enough for you to get the diesels going.
How much time is there before a meltdown?
It depends on the plant. It depends on whether it's a boiling-water reactor or a pressurized-water reactor. Basically, [in both] you have the benefit of natural forces such as convection. There is a coolant loop no matter what, so you end up to some degree cooling the core because the heated water rises and colder water gets pulled in. But that's not as effective as a pump bringing in cool water. Just to speak very broadly, you have many hours to restore power to the system to get normal cooling going. It's really not possible to get more specific than "many hours."
But generally, less than 24 hours.
That's fair to say.
What's the worst-case scenario?
The event we are looking to avoid is damaging the core. Once you start damaging the core, you are then releasing radioactive material into the coolant and thereby increasing the chances that something travels outside the reactor.
The reactor that was not cooling properly in Japan, the Fukushima Daiichi No. 1 reactor, was a boiling-water type. How are these different from pressurized-water reactors in terms of cooling?
Particularly useful to boiling-water reactors is a system that is steam driven. It does not require an outside power source. Steam generated by the heat of a cooling down reactor has enough force to run a turbine, which then runs a pump that provides coolant to the core. That sort of system is supposed to withstand an earthquake, and that can run for an extended period. It's a self-limiting condition. That system does use batteries for the controls, but it can also be operated manually. So even in the face of a complete station blackout—you don't have any power at all—there are methods for using the steam-driven pump to continue to keep cooling going.
Are there other coolants besides water?
Water would be it, essentially. The reserve tanks at a reactor contain the same grade of water in terms of purity and chemical composition that are normally used in the core. It is possible, if you have a situation where you have exhausted that source of coolant to introduce, quote–unquote, regular water. That will do the job of cooling.
Why do nuclear power plants need electricity to be cooled?
Nuclear reactors are net positive in terms of supporting the grid. They produce much more electricity than they need to run their systems. As a basic design feature in the U.S., plants are not literally self-powering. That's by design, because you don't want to end up in a situation where a problem at the plant cuts off its own power source. Therefore, the primary means of power for a plant in order for it to run is electricity from the grid. As a general matter, for U.S. plants, if you can't use power from the grid, you shut down.
What kinds of events could knock out a diesel generator?
You always have the possibility of just plain old failure. That's why you have multiple diesels at a plant for redundancy's sake. It can be the case that diesel itself is running properly but the distribution system, the buses or the cabling could be misaligned to the point where the diesel detects that its power is not being accepted by the plant. It's not going to run if it's trying to generate power and that power's not going anywhere. When we say a diesel fails, it's not always a problem with the diesel [itself].
How long does it take to cool down a reactor?
There are design specific variables there. The easiest way to answer that question is that NRC regulatory requirements for emergency power supplies is that they be available on the order of a month. You can render a plant in an acceptable condition within a few hours. However, heat is still being generated. If you had to stop, at any point, carrying away that heat, it would start building up again. Emergency cooling systems have to be available for weeks.
