EP09 - Neutronics Simulations (to design safe reactors)

Discover how Monte Carlo neutronics simulations model neutron behavior to design safe reactors and shielding. Learn the computational physics behind criticality safety calculations and why probabilistic simulation is the only practical tool for reactor design.

Tweet coming soon

I've got it, Kurumi. I know how to design the shield. I just have to guess where the neutrons bounce!

You're guessing? With radiation?

Dang it.

But if I launch it *exactly* the same way again... it should go to the same place, right? Newtonian Physics!

Neutrons aren't pinballs, Shez. They are **Quantum Particles**.

They don't follow Newton. They follow **Probability**.

Imagine shooting a million balls at once. And every bumper has a different probability of eating the ball or spitting it out.

You can't predict *one* grain of sand. But you can predict the **Pile**.

Welcome to the **Monte Carlo Simulation**.

Like the casino?

We trace the life of one neutron.

Step 1: It is born in fission.

Step 2: It travels a distance. How far?

It travels 10cm. Why? Because the **Mean Free Path** in water is roughly that long.

So you roll dice to determine movement?

Step 3: What happens? Does it Scatter? Get Absorbed? Or Fission?

We roll the dice again.

Result: **SCATTER**.

We do this over and over until the neutron dies (absorbed) or escapes the reactor.

That's... tedious. That's just one neutron.

One neutron is a "Random Walk" (A Drunkard's Walk). It's noise.

But **One Billion Neutrons**?

The chaos converges into a pattern.

The Law of Large Numbers takes over.

This is a **Flux Tally**.

We didn't solve a differential equation. We just rolled the dice a billion times and counted where they landed.

It's beautiful. It looks like a weather map.

It's **MCNP**. Monte Carlo N-Particle code. We use it to predict the invisible.

Can we simulate my design? I put a concrete wall here.

Let's run it.

Launching 10 million particles.

**RUNNING... NPS (Number of Particles) = 10,000,000**.

Look! The concrete wall stops them! It's blue behind the wall! I'm safe!

Are you? Look at the **Streaming**.

Neutrons act like a gas. They find the path of least resistance. You left a gap. In the real world, that beam would be a death ray.

I thought it would just... shadow it.

This is why we simulate.

We want to find the death rays before we build the concrete.

And look at your K-effective.

1.05? Is that good? It's close to 1.

1.05 is **Supercritical**.

Your reactor power is doubling every minute.

In 10 minutes, your garage melts.

How do I fix it?!

Add more Boron. Or remove fuel.

You need k = 1.0000. Exactly.

Adding control rods...

Now it's under 1! But it won't start!

It's **Subcritical**. It's a paperweight.

This is impossible! It has to be perfect!

That is Nuclear Engineering.

We use this data. These are the probabilities.

See this spike? That's a **Resonance**.

At that specific speed, Uranium gets very hungry for neutrons.

The simulation knows these probabilities for every isotope in the periodic table.

We did it. We simulated a stable core.

So... now can I build it?

You have the **Neutronics** right.

But you forgot one thing.

What?

In the simulation, materials don't melt. In reality, heat expands metal. Water boils. Pipes burst.

You solved the *Physics*. Now you need **Thermal Hydraulics**.

There's always another simulation!

Now, back to the one variable you forgot. **The Heat.**

Wait! The neutron count is stable! Why is it melting?!

Neutrons create heat. Heat needs to leave.

If the heat is generated faster than the water can carry it away...

**Meltdown.**

You have a "Neutronically Stable" puddle of lava.

But... the water is right there! Why didn't it cool the rod?

Because you boiled it too hard. Enter **Critical Heat Flux (CHF)**. Let me show you a metaphor.

It's floating!

That is **the Leidenfrost Effect.**

The water boils so fast it creates a layer of steam underneath itself.

Steam is a gas. Gas is an **Insulator**.

Look at your fuel rod.

The heat is trapped inside the rod. It has nowhere to go. So the temperature shoots to infinity.

So bubbles are bad?

Small bubbles are good (**Nucleate Boiling**). They mix the water.

A wall of bubbles is death (**Departure from Nucleate Boiling - DNB**).

How do I pop the bubbles?!

**Turbulence**.

You need to mix the water violently.

This is **Mixing Vanes**. They swirl the water like a tornado.

The centrifugal force throws the heavy water onto the hot rod and pushes the light steam away.

It worked! The heat is transferring!

Now for the final boss. **Coupling**.

Coupling?

Physics isn't separate. Neutrons affect Heat. But Heat affects Neutrons.

When Uranium gets hot, the atoms vibrate faster.

This makes the U-238 "fatter." It absorbs more neutrons.

This is the **Negative Temperature Coefficient**.

If the reactor gets too hot, the physics naturally kills the chain reaction.

Finally, green lights across the board.

I thought Nuclear Engineering was just splitting atoms.

But it's half Quantum Mechanics, half Plumbing.

Maybe **Quantum Plumbing**?

You simulated the Neutrons (MCNP). You simulated the Fluid (CFD). The computer says it won't melt.

Now we just have to hope the welder isn't drunk when he joined the pipes.

Yeah... hope.