Why is external power loss such a danger for nuclear power plants? Can’t they use power that the plant generates for pumps and other mission-critical systems?


If I’m not mistaken, external power loss caused the Fukushima disaster. Also, as far as HBO told me, they were testing this very scenario when the Chernobyl nuclear power plant exploded.

In: Engineering

The reactions in a nuclear core are really complex. They have to be balanced on a knife edge. When everything is balanced they are self regulating, but a sudden swing and they run away very quickly because there is a time lag of up to 6 hours in the production of neutrinos in the nuclear reaction.

You suddenly drop regulation- coolant for instance, you’re potentially running with a thermal power output from 6 hours ago with no where for the heat to go. This is what happened- simply put- in Chernobyl.

There’s a guy on YouTube called Scot Manley. He does rocket and space stuff but he recently did a very nicely explained tech low down about what happened in Chernobyl.

So, a nuclear power plant *can* power itself if it was designed to do so, but there’s an added issue that makes the ability for the reactor to self-power kind of redundant, which makes it an unnecessary cost for an already expensive project.

In the event of a power failure, the (broken) energy grid essentially becomes a kind of endless pit that the reactor keeps dumping energy into, but nothing comes back out in order to power the reactors systems, and it actually takes power to run the systems that end up disconnecting you from the grid. Thus, to have a plant that can power its own systems, you have to be able to detect a grid failure, throttle the reactor, vent excess steam, and disconnect from the grid, all essentially automatically and within a few moments, and you can’t afford to miss that window.

The key problem is if the reactor is scrammed; if the control rods are in the reactor, the reactor cannot generate steam, which means it cannot generate power, which means that even if the reactor could power itself, it’s no longer generating enough power to do so, ergo you need *something* to power the systems in the event of a grid failure **and** a reactor scram. Hence, you’ll install diesel generators somewhere near the reactor.

However, if you’re already going to install those reactors, it’s not really worth it to make the reactor power itself; the diesels can do it if the grid goes down, provided they can spool up quickly enough to provide thorough control of the reactor and it’s systems.

The problem at Chernobyl is that the diesels would take about 60 seconds to spool up, hence the reason for the underlying test was to figure out if the reactor could be run by bleeding off the angular momentum of the turbines.

At Fukushima, the problem was the placement of the diesels; large gensets are often liquid cooled, and so you’ll want something to dump that heat into. The ocean itself is *very* good at this, but that also necessitated placing the gensets close to the ocean in order to keep costs low. Hence when the Tsunami hit, the generators were flooded, and the plant went into a meltdown. Had the generators been located on the other side of the plant, further uphill, they would have been fine, and the disaster never would have happened.

Electricity behaves very much like a physical substance. When you put electricity into an electrical system it has to go *somewhere.* If it doesn’t have anywhere to go, then it turns into heat. This means that if you have a lot more electricity going into the system than the system can use, you can generate so much heat that the system itself catches on fire – so all of the electrical transformers will explode, the wires will melt, and eventually your electrical turbines will catch on fire.

Nuclear power plants are typically designed to generate a constant, massive amount of electricity. This is done by running all of the steam that the reactor generates through a handful of turbines. The plant can only scale how much power its generating by shutting off turbines, but it can’t control the power output of any individual turbine – each turbine is either running at 100% or 0%. This means that nuclear plants have a minimum power output equal to 100% of the output of 1 of their turbines – which in most plants is going to be enough power to run a small city.

Running the pumps to keep a reactor cool takes a tiny amount of power. When a nuclear plant loses external power, the problem isn’t that it can’t accept power from the grid, but rather than it can’t send excess power to the grid. The minimum amount of power that the plant can generate dwarfs what it needs to run its cooling pumps, and since it can’t send its excess power to the grid it ends up with a huge power imbalance. If the plant tries to run its pumps under those circumstances all that will happen is that all of the infrastructure in the plant will overload and destroy itself. Because of that, when a nuclear plant loses its external connection the only safe thing it can do is to fully shut down its turbines.

That doesn’t mean that nuclear plants couldn’t be designed with a small turbine that was only used to run the pumps, or to be designed with a way to bleed off steam to allow its main turbines to be run at less power. But both of those add a lot of cost to an already expensive project to solve a problem for which much cheaper solutions are available. They also don’t necessarily add any safety – for example, Fukushima had diesel backup generators to run the pumps in the case of an external power loss.

The problem at Fukushima was that the diesel backups also sustained damage that prevented them from being used. Which is sort of the problem with having a reactor driven cooling system – anything that damages the reactor is probably going to also damage your backup system. Modern reactor design focuses on making it so that reactors are safe even if they lose active cooling, rather than on figuring out ways to maintain the active cooling.

Normally, yes, a nuclear power plant can be powered by the electricity it generates itself, but there are many different emergency situations where you have to shut down the reactor by inserting the control rods (frequently called a reactor scram, or AZ-5 in Chernobyl). This means the core is no longer generating power, and the turbine is no longer spinning. Even after the reaction has been shut down, the fuel still needs constant cooling, which means if the plant itself isn’t generating power, you need an external power source for the coolant systems. Usually this is accomplished with diesel generators.

At Fukushima, the reactors were scrammed after the Earthquake was detected, as was normal operating procedure. The problem was that the plant’s seawall was not designed for a tsunami the size of the one generated by the earthquake, and seawater flooded the area. The diesel generators were also poorly located on the lowest level of the plant. They were inundated with seawater and unable to operate. This left the core with no cooling, which resulted in core meltdowns.

Chernobyl also had backup diesel generators in event of a power loss. The test they were doing was to see if there was power generating by the spinning down turbine to power the cooling system in the 1 minute it would take for the diesel system to come to full power.