Chernobyl Accident — Visually explained

The worst nuclear accident in history, the Chernobyl disaster, why did it happen?

Today I’ll show you. Step by step, I’ll explain the relevant nuclear physics. I’ll implement that in a code and make a very basic model of the Chernobyl reactor. Then, in the second part of the video, I’ll use the model to explain and redo the same terrible decisions in the Chernobyl control room, and you will be able to see how the reactor responds — and exactly what went wrong.

Prefer to watch? This article is available as a video format as well:

A very simple reactor

But let’s start with the basics. In any nuclear reactor, essentially two things happens: Either you bring up the reactivity or you bring it down. You must bring the reactor up to power, and keep it balanced there. The power from the nuclear reaction heats up water into steam, which then drives a turbine to generate power.

This is done by fission.

I recommend watching my video about critical mass where I simulate a nuclear reactor and also an atomic bomb while going into deeper details about fission. Video is here:

But don’t worry, I’ll assume you are lazy and quickly recap now, while also building up the model to simulate the Chernobyl reactor.

The nuclear reactor uses Uranium of the isotope 235. This element has the property of undergoing induced fission which is when the element absorbs a neutron. The element will split. i.e. fission into to 2 new lighter daughter elements, and release 3 new neutrons:

This releases about a million times more energy than if a chemical reaction occurred such as a TNT-molecule reacting.

Now, these daughter elements are not important for the bigger picture. So not to confuse ourselves visually in the model, instead of showing two daughter elements, I’ll simply just show whether the element is Uranium235 or not.

We can stack many Uranium235 in a metal grid and fire a single neutron.

We just created a nuclear chain reaction. This is what we want in a nuclear reactor, although a bit more controlled, and not this explosive.

In the reactor we will have different isotopes of Uranium in the fuel, and only a few percentages is the fissile type, Uranium235. This is how the metal is extracted from the ground. Having highly enriched Uranium235 in a reactor is not only extremely expensive but also too dangerous.

You can see this reaction seems way more stable. The goal is to get the reactor up to some amount of reactions per second, and once you are there, you keep it stable by ensuring each loose neutron on average only hits a single Uranium235 nucleon. By doing that, the reaction is neither increasing nor decreasing.

I want to add two more things to the model before continuing — these are just for helping my model and not really of importance in actual reactors. The first thing I want to add, is the ability to replace used up fuel. Since I only have around a thousand simulated Uranium nuclei I need to be able to magically pop in new enriched Uranium:

The second thing I want to add is in the model is the radioactive decay: Spontaneous neutron emission. This is when a radioactive element suddenly releases a neutron — a naturally occurring process. I want to add it because there is a chance all neutrons in the model disappear, and I don’t want to restart the animation. Let’s give the property to all non-Uranium nuclei i.e. the gray circles.

Control rods

Alright, now I want to build a model of a nuclear power plant. And specifically, I want to simulate the same type of reactor used in Chernobyl — the RBMK reactor which apparently is an abbreviation for Ridiculously Badly Made Kettle [Citation needed].

The first thing I want to add is control rods. Control rods have the property to absorb neutrons.

Alright, we don’t have everything yet to simulate the RBMK reactor but for now let’s run what have. It’s a very basic nuclear reactor model. Normally, reactor reactivity is measured in megawatts, but I’ll just count the number of neutrons present in the reactor. You can see that number on top. I’ve coded some very basic automatic control of the control rods. Every second rod will go up if there is under 40 neutrons inside the reactor. If there is above, the control rods will be inserted instead.

As you can see, by SLOWLY pulling out control rods, I am able to increase reactivity, whereas re-inserting them decreases reactivity.

Just to state the obvious, you can cool down a reactor by inserting rods and heat up a reactor by removing the rods, allowing neutrons to flow freely. This gives operators control of the reaction.

Water

Alright, let’s keep adding real physics concepts to the model and build up a more realistic model. As I briefly mentioned, the reactor is surrounded by water, which heats up and drives a turbine.

However, water has a low chance of absorbing a neutron. I’ll depict water as a small box.

Whenever a neutron scatters with water, it will transmit kinetic energy heating up the water. In the model whenever a neutron is inside the water I’ll heat it up. When water is cold I’m coloring it blue, and then fading into red as it gets hotter.

But getting too hot means the water evaporates. When this happens the water is gone, and can no longer absorb neutrons. In my model this is going from red to completely gone. When the water cools of it condensates and reappears.

The front of the neutron burst is mostly absorbed while heating up the water. This means the backside can easily propagate without being absorbed. In reality of course there would be some temperature exchange between each water block but let’s just ignore that fact.

So for our reactor that means presence of a lot of water will keep the reactivity down whereas heating up the water will lead to holes or voids, and the reactivity will go up even further. Reactors with this property are said to have a positive void coefficient.

Let’s see that in action from the same example before with the control rods. Notice that we now have to lift the control rods up even further, because the water is helping to absorb the neutrons.

But because the control rods keep the reactivity down to 40, you won’t see enough neutrons to evaporate the water for now. You will have to wait a bit to see the positive void coefficient in action.

Decay chains / Xenon

We only need two more things to simulate the Chernobyl accident.

For this chapter, I’ll keep it short and concise. After absorbing a neutron, at a later time the nuclei will have a chance to translation into the isotope of Xenon135. If you want to know why and how, go look about Decay chains and decay branches.

Xenon has the property to strongly absorb neutrons. This is important for the Chernobyl ancient as you can effectively poison your reactor, i.e. make it unable to go up in power, because most loose neutrons just gets absorbed by Xenon instead of undergoing fission.

In my model all elements that have undergone fission now has a chance of becoming Xenon at a later stage. I’ll denote Xenon with the same dark gray color as the control rods. Once it’s a Xenon element, the only way to undo that is for the element to absorb a neutron. This is what is to referred to as burning away Xenon.

Now let’s run the model again with the Xenon property.

You will see once again now it’s harder for the reactor, and we must raise the control rods even more to get the same reactivity level around 40 neutrons.

Moderation

Last thing to add!

When neutrons are released from the nuclei, they fly out with around 5% the speed of light. This speed is way too fast for fission absorption. It turns out the chance of interaction is incredibly low. This is called the nuclear cross-section, it can be thought of as the probability of a reaction. Not exactly a probability though because it’s measured in the unit barn and not percentage.

Let’s add that property to the model. Now neutrons sent from nuclei are fast, I will denote that with a white dot in the neutrons. When they are going fast, I’ll set the interaction chance of 0%.

For fission, you must first use a so-called moderator. Something for the neutron to scatter on, and absorb some kinetic energy. In the RBMK reactor, all fuel rods were surrounded by graphite for moderation — yes the same thing as in your pencils. In the model I’ll depict a moderator as white rod with a gray border.

After collision with a moderator, the neutrons are no longer too fast and are instead called thermal neutrons — which basically just means slow neutrons.

Let’s have these thermal neutrons behave just as before, with an interaction chance of 100% when the neutron a nuclei touch. This is of course not correct, but you know kinda. When thermal neutrons collide with the moderator, we will just let them fly right through.

A bit confusing I know, to make sure we are on the same page I’ll run the model without controls rods, water or the xenon property. Only with the fission fuel and moderators slowly being inserted. Hopefully this depicts the influence of moderation and nuclear cross-section.

This is yet another control of reactivity, removing moderation means neutrons will have a much lower chance of undergoing fission, while more moderation means more neutrons will have a high probability of fission.

Finally, let’s see everything together. A lot is going on, so for this model I added a legend of all the different elements below. I know there is a lot going on.

Now we are ready to see what went wrong in Chernobyl.

The day of the accident

Event 1

I will recreate the same events as in the control. Here is the first one. Event 1: Reactor normal.

The simulation I am showed you just before (the ones with a legend of all elements. Link here) Is showing how the reactor was running a stabile just before the accident. It was running at around 50% power to meet power demands for the grid line. The Xenon was being burned off just as quickly it is being made — reactor is stable. We can pretend the 50% power corresponds to 40 active neutrons, which is what the control rods here is trying to stay stabile at.

To illustrate the state we just saw, here I am plotting the reactivity of the reactor according to the events. This event it was at 50%.

And on the other plot, I’ll show how many of the control rod was in. It was a few. How much is Xenon — not much because it was being burned away straight away. I’ll also plot how much water in the reactor, The water was stable cooling the reactor down. were in and some were out to match the 50% power. The water i.e. coolant was running normally with some voids here and there.

Event 2

Day of the accident, Event 2: Power reduction.

Let’s continue the simulation from where we stopped last time. We will do that from now on.

What happened then? A safety test was scheduled later this night. This test required the power to run at reduced at around 30% power. We can pretend 30% power corresponds to 20 active neutrons and set the automatic control rods to try and hit this number. What we can see here, is the Xenon is still being generated from a delayed reaction when the reactor ran at 50% power before, but now only 30% power burns it away.

Let’s see the state of the simulation we just saw

Control rods were inserted in order to try and reach this lower power level. In contrast they increased the water flow cooling. However the raise in control rods caused lower power and not as many neutrons are present to burn of the Xenon.

Let’s see it in action, continuing from where we left last time. We will do that from now on.

Event 3

Day of the accident, Event 3: Power drop.

After that, the power unexpectedly dropped to 1%. Because of the xenon buildup we are sometimes reaching very low levels.

So the power level is plummeted to 1%. They didn’t change anything to the control rods of course, the reactor was simply stalling because of the extra high water flow with no voids and the Xenon poising continually building up from the reactor running at 50% power.

Let’s see that.

EVENT 4

Day of the accident, Event 4: Power up attempted.

Now many safety systems were turned off, safety protocols ignored, automatic control of control rods was turned off In an attempt to power up reactor.

They were only able to get the reactor running at 7% by pretty much removing all control rods.

Notice the control rods are raised, but the water and xenon keeps the reaction In check.

So the state is 7% power, all control rods raised, Water level is still cooling the core and Xenon keeps building up from the delayed reaction.

Let’s see that.

EVENT 5

Day of the accident, Event 5: Test starts”.

At still stuck at 7% power the computer warns them to shut down immediately, so they turned off the computer instead. They could not get the power higher than 7% And started the test.

Pretty much all control rods are still raised, The test included switching off half of the recirculation pumps temporary. To reflect the switching off half of the recirculation pumps temporary, I’ll simply allow the water to cool of much slower. The effect is more water boils away, causing more xenon to off. This is a positive feedback loop and suddenly much of the water boils away, and all the free neutrons burns away the Xenon.

Right, so the reactor is stuck in 7% power. Water cooling is increadbly low becuase of the test. Voids begin to form causing a higher neutron flux, causing the Xenon to burn away.

The reactor in is set in such a crazy state..

EVENT 6

Day of the accident, Event 6: SCRAM”.

Here is something I hadn’t added to the simulation: The control rods had Graphite displacer, commonly said “Graphite tips” hanging at the end and not the full length of the reactor. Looking like this. So actually, This whole time, The scenario should looked like this instead. I’ll add that to the simulation

Let’s continue the simulation with these added changes

Suddenly, the reactivity increases unexeptedly fast in the reactor and operators pushed the SCRAM button aka AZ-5 a safty switch.

The scram button job is to put in all the control rods as fast as possible. It was a lifeline they always thought they had in case something vent wrong in this test.

Anyway, pressing this SCRAM button takes some time to be deploy the control rods. They are slow! The operator didn’t know what they just had done. Normally the top had a lot more stopping power than the accelerating bottom graphite. But the reactor was in such a unstable state, that the bottom had a much much stronger accelerating power. As a result, the SCRAM actually increased the reaction rate in the lower half of the core. After a little, the control rods and graphite displacers were stuck — they might have just melted together to the core. Now moderator was now stuck endlessly accelerting the reactor. Nothing was there to stop the reaction.

It’s just all gas and no breakes. In combination with the effect from the SCRAM effect, more voids burns away more xenon wich burns away more water and soon. This is the state:

350 Kg or 400 pounds of metal rods jump up and down due to the extreme steam pressure. Power running over 100000% it’s design capacity.

Explosion occurred. See it here:

Conclusion / end

So hopefully now you have an intiition on what vent wrong in Chernobyl and also learned a little nuclear physics. I basicly only covered the physics and the incredible dangerous design of The RBMK reactor, but there are tons of other factors that also played a role, which we didn’t cover.

I don’t want you to walk away with a feeling nuclear reactors are unsafe.

This is just one reactor design, and many much more safe reactor design exist, for example ones with a negative void coefficient. I could simulate that for you one day. They are impossible blow up.

Watch all the events from start to finish here:

Thanks for reading!