Nuclear Magic: Where's the Throttle?

Figure 1. A U-238 resonance at an energy level of 6.67 eV. Vertical scale is likelihood of absorption. At low temps only the few neutrons right around 6.67 eV can be absorbed. As the U-238 heats up, more and more neutrons are candidates for absorption

There is a widespread belief that nuclear reactors cannot loadfollow. This is nonsense. In most countries, loadfollowing is a grid requirement. In fact, stable reactors are inherently loadfollowers. A stable reactor is one whose power decreases as the reactor temperature increases. This is the property of any well-designed nuclear reactor, and a strongly enforced, legal requirement.

A few early designs could be put in a state where the opposite was true, an increase in core temperature increased power which increased temperature which increased power, which ... and the whole thing goes boom. One such design was the Russian RBMK at Chernobyl. You had to work pretty hard to put the RBMK in such an unstable state; but human ingenuity was up to it, and we got the Chernobyl explosion.

All temperature stable nuclear reactors have a remarkable ability. They automatically adjust their power output to the load. If the electric load increases, the turbine control system will speed the steam flow up. The increased steam flow will extract more heat from the steam generator. This will decrease the temperature of the reactor coolant. The reactor core temperature will drop, and the reactor will increase power, without us doing anything. It's built into the physics.

Conversely, if the load drops, temperature increases, and reactor power decreases.¹ If the loads drops completely, say due to a station black out, the reactor temperature will shoot up, and the reactor will shut itself down. The shutdown does not depend on a system that senses a problem and then tries to react to it. There is nothing a confused operator or malfunctioning control system can do to prevent this process.

For an old tanker operator, it's magic. For the longest time, I kept asking: where's the fuel throttle? There is none. Weird, and if you think about it, wonderful.

How is this possible?

The few inquisitive minds in the choir will be asking: how does this work?

Reactor control is all about neutron balance. A neutron can end up either

1) escaping the reactor core,

2) being absorbed into a non-fissile nucleus, such as U-238.

3) or causing a fission, by hitting a fissile nucleus such as U-235.

The main way the reactor is controlled is by changing the absorbed versus fission fractions.

When the core heats up, for example because the load has decreased and the reactor is putting out more power than we are taking out of it, a lot of things happen.

a) All the components in the core expand.

This can have either a positive or negative effect on power output depending on the component and design. For example, in a water moderated reactor the decrease in water density with increase in temperature reduces the amount of moderator in the core. Less moderator means less fission. In a well designed reactor, the overall effect of expansion needs to be negative or nearly so.

b) The Doppler effect. This is the biggie. It is not all that easy to understand. We need to get into the weeds. So here we go.

In order to decrease reactivity, we want more of the neutrons to be absorbed rather than fission. If a neutron hits an atom of say U-238, it will only be absorbed if the neutron's energy (aka speed) matches the difference in energy between the ground state of U-238 and one of the excited states of U-238. This is called a resonance.

If the U-238 atoms were stationary, the resonance peaks would be very narrow. Only the neutrons whose energy closely matched the difference between the states could get absorbed. You would get very little absorption. But the U-238 atoms are all bouncing around in Brownian motion. What counts is the relative speed of the neutron to the speed of the U-238 atoms. There are now a lot more ways for the neutron's energy (as seen from the point of view of the U-238 atom) to match the difference in the energy of the two states. The resonance peak is effectively broadened, Figure 1. More neutrons are absorbed which means less neutrons to fission.

If we now increase the temperature further, the U-238 atoms will bounce around even faster, all the resonance peaks further broaden, still more neutrons get absorbed, and the reactivity and power goes down. If the temperature rises still further, you finally get to the point where the reactivity has dropped below critical, and the reactor shuts itself down.

If nobody does anything, the shutdown core will eventually cool to the point where the narrowing of the resonance peaks will increase reactivity enough to start the reactor back up again. The cycle will then repeat until the operator or the control system steps in. This is called simmer mode, and, like the decay heat, must be allowed for. In the designs that I am familiar with, simmer mode will average about 1% of normal power.

Importantly the Doppler effect, unlike much of the expansion effect, is very fast.

An important property in all this is temperature margin. If we are going to rely on temperature changes to control the reactor, the reactor must be able to handle the resulting temperature fluctuations without ill effect. Few machines tolerate repeated rapid changes in temperature well. In many cases, including the Light Water Reactor, this dictates some form of active control even though the reactor is stable. But the stability guarantees that we won't have a run away power excursion even if the active control fails.

1. Light water reactors can't take full advantage of this magic. The temperature of boiling water reactors is stuck at the saturation temperature. Pressurized water reactors must keep the coolant temperature within very narrow limits to avoid boiling. They need some form of active control.