Non-intrusive Monitoring of Sealed Microreactor Cores Using Physics-Informed Muon Scattering Tomography With Momentum Measurements

This paper introduces $\mu$TRec, a physics-informed muon scattering tomography framework that significantly enhances the detection of missing fuel in sealed microreactor cores by reconstructing curved muon trajectories and incorporating momentum measurements, thereby outperforming conventional methods like PoCA in both sensitivity and speed under realistic cosmic-ray conditions.

The Gist

Imagine you have a brand-new, high-tech car engine that is completely sealed inside a thick, unbreakable metal box. You can't open the box, you can't take it apart, and you can't ask the driver what's inside. However, you need to know: Is the engine complete? Did someone steal a crucial part?

This is the exact problem scientists are facing with microreactors—tiny, portable nuclear power plants designed to be shipped anywhere and run with very few people on-site. Because they are sealed and compact, traditional ways of checking them (like opening the lid or counting fuel rods) don't work.

This paper introduces a clever new "X-ray vision" technique called Muon Scattering Tomography to solve this mystery. Here is how it works, explained simply:

1. The Invisible Detectives: Cosmic Rays

Instead of using a giant machine to shoot X-rays at the reactor, the scientists use muons.

• What are they? Muons are tiny particles created when cosmic rays from deep space hit our atmosphere. They rain down on Earth constantly, like invisible rain.
• Why are they special? They are like "ghosts" that can walk through thick walls, lead, and dense metal that would stop normal light or X-rays. They can pass right through the sealed microreactor.

2. The Old Way vs. The New Way

When these muon "ghosts" pass through the reactor, they don't just go straight. If they hit a heavy piece of fuel, they get bumped and change direction slightly (a bit like a pinball hitting a bumper).

• The Old Method (PoCA): Imagine trying to guess where a pinball hit a bumper by drawing a straight line from where it came in and where it went out. You guess the "closest point" where those lines might have crossed. It's a rough guess, like trying to find a needle in a haystack by looking at the shadow. It often misses small details or gets confused by the complex shape of the reactor.
• The New Method (µTRec): This is the paper's big innovation. Instead of guessing a straight line, the scientists use a physics-based map. They know exactly how heavy muons are and how they behave when they hit different materials.
  • The Analogy: Think of it like a detective who doesn't just guess where a suspect ran, but uses the suspect's speed, the wind, and the terrain to calculate the exact, curved path they took.
  • The Secret Sauce: The new method uses momentum (how fast and heavy the muon is). If a muon is moving fast, it's harder to bump. If it's slow, it bounces easily. By knowing the speed of every single muon, the computer can draw a much more accurate picture of the inside.

3. The "Missing Flake" Test

To prove this works, the researchers built a digital model of a microreactor (based on a real design called eVinci). It's a honeycomb of 61 fuel pieces.

• The Challenge: They secretly removed one fuel piece (a "missing flake") to see if the system could find it.
• The Result:
  • Using the old method, the missing piece was hard to see, especially if the "picture" wasn't super sharp.
  • Using the new method (with momentum), the missing piece popped out clearly, even with fewer muons. It was like switching from a blurry, grainy photo to a crisp, high-definition image.

4. Why This Matters

• Speed: Because the new method is so much better at finding the truth, you don't need to wait as long for enough muons to pass through. You can verify the reactor in a fraction of the time.
• Real-World Use: The team tested this with two types of "rain":
  1. Laser beams: A controlled, perfect stream of muons (like a laser pointer).
  2. Cosmic rays: The natural, messy rain of particles from space.
  • The Good News: The new method worked great in both scenarios. Even with the messy natural rain, it could spot the missing fuel.
• Robustness: They also tested it with "blurry" detectors (simulating cheaper, less perfect equipment). Even with some blur, the system still found the missing fuel. This means it could actually be built and used in the real world without needing million-dollar, perfect sensors.

The Bottom Line

This paper presents a new "super-spy" tool for nuclear safety. By using the natural rain of cosmic particles and a smarter math algorithm that understands physics, we can look inside a sealed, tiny nuclear reactor and instantly spot if a piece of fuel is missing—without ever opening the box.

It turns a complex, invisible problem into a clear picture, ensuring these future power plants are safe, secure, and working exactly as they should.

Technical Summary

Here is a detailed technical summary of the paper "Non-intrusive Monitoring of Sealed Microreactor Cores Using Physics-Informed Muon Scattering Tomography With Momentum Measurements."

1. Problem Statement

Next-generation microreactors (1–20 MWe) offer portable, sealed, and semi-autonomous power solutions for remote and critical facilities. However, their compact, highly heterogeneous internal geometries (containing fuel flakes, control drums, heat pipes, and reflectors) and sealed nature pose significant challenges for traditional nuclear safeguards.

• Limitations of Current Methods: Conventional verification relies on access to declared materials, containment/surveillance, or bulk accountancy. These are ineffective for sealed microreactors where direct inspection is limited, and internal anomalies (e.g., missing fuel, unauthorized modifications) are localized and difficult to detect via bulk measurements.
• The Gap: There is a need for non-intrusive monitoring techniques capable of verifying core integrity and detecting localized defects without opening the reactor or relying on operator declarations.

2. Methodology

The authors propose and evaluate µTRec, a physics-informed muon scattering tomography framework designed to reconstruct 3D internal structures using cosmic-ray or laser-driven muons.

A. Simulation Setup

• Target Geometry: A proxy model of the eVinci heat-pipe microreactor (Westinghouse concept). The core is hexagonal, containing 61 fuel flakes (UO2), 12 rotating control drums, 7 shutdown rods, and embedded heat pipes.
• Muon Sources:
  1. Idealized: Laser-driven parallel beam (5 GeV).
  2. Realistic: Cosmic-ray muons (0–60 GeV, zenith angle 0–90°) simulated via GEANT4.
• Detector Configuration: A four-plane scintillation detector array (upstream/downstream pairs) to track incoming and outgoing muon trajectories.

B. Reconstruction Algorithms

The study compares two reconstruction approaches:

1. PoCA (Point of Closest Approach): A traditional geometric method assuming a single scattering event at the intersection of incoming and outgoing tracks. It ignores momentum and treats the path as a straight line with a single deflection point.
2. µTRec (Physics-Informed):
   • Physics Model: Explicitly models Multiple Coulomb Scattering (MCS) using a Gaussian probability distribution rather than a single point.
   • Trajectory Estimation: Uses a Bayesian framework to estimate the most probable curved trajectory of the muon through the volume, constrained by measured entry/exit tracks.
   • Momentum Integration: Incorporates per-muon momentum ($p$) and energy loss ($E_{in} - E_{out}$) into the scattering model. The scattering angle variance is inversely proportional to momentum squared ($\sigma_\theta \propto 1/p^2$).
   • M-Value Mapping: Instead of mapping raw scattering angles, the algorithm calculates an M-value for each voxel:
     $$M(p, \theta) = \log_{10}\left( (\theta \cdot p)^{2.4} \right)$$
     This metric normalizes scattering by momentum, improving density estimation accuracy.

C. Evaluation Metrics

• Detectability ($D_{P_{means}}$): A statistical metric based on the separation of mean voxel values between intact and missing-fuel regions, normalized by variance.
• Scenarios: Tested under varying voxel sizes (4–50 mm), muon counts (0.5M–3M), detector resolutions (5–10 mm spatial, 5–10% energy), and spectrometer configurations (0, 1, or 2 spectrometers).

3. Key Contributions

1. First Application to Microreactors: Extends muon tomography from cargo/spent-fuel screening to the specific, complex geometry of sealed microreactors.
2. Physics-Informed Trajectory Modeling: Demonstrates that modeling the muon path as a probabilistic curve (via MCS) significantly outperforms straight-line approximations (PoCA) in heterogeneous media.
3. Momentum-Informed Reconstruction: Quantifies the specific benefit of incorporating per-muon momentum data (via spectrometers) into the reconstruction algorithm, showing it drastically improves defect identification.
4. Robustness Analysis: Validates the method against realistic detector constraints (finite spatial and energy resolution) and varying instrumentation costs (single vs. dual spectrometers).

4. Key Results

A. Impact of Momentum Information

Incorporating momentum data significantly enhances detectability ($D_P$):

• Laser Source (5 GeV): Momentum-informed µTRec improved detectability by 65% to 149% across voxel sizes compared to momentum-agnostic reconstruction.
• Cosmic-Ray Source: Improvements ranged from 18% to 105%. The benefit was most pronounced at finer voxel sizes (4–10 mm), where accurate trajectory assignment is critical.
• Visual Quality: Momentum-informed reconstructions clearly resolved the missing fuel flake, control drums, and shutdown rods even at 50 mm voxel resolution, whereas momentum-agnostic methods failed to identify the defect reliably at larger voxel sizes.

B. µTRec vs. PoCA

µTRec vastly outperformed the standard PoCA algorithm:

• Detectability Gains: In the momentum-informed case, µTRec achieved 326% to 392% higher detectability than PoCA for equivalent muon counts.
• Low-Count Performance: µTRec could identify defects with as few as 0.5 million muons (reducing acquisition time), whereas PoCA failed to resolve defects at this count, even with momentum data.

C. Detector and Instrumentation Constraints

• Resolution: The method is robust to realistic detector limitations. With 10 mm spatial resolution and 10% energy resolution, detectability dropped by only 8.88% compared to ideal conditions.
• Spectrometer Count:
  • 0 Spectrometers: Baseline performance.
  • 1 Spectrometer (Entry only): Increased detectability by 101%.
  • 2 Spectrometers (Entry + Exit): Increased detectability by 105%.
  • Conclusion: The marginal gain from a second spectrometer (4%) suggests a single-spectrometer configuration offers the optimal cost-performance trade-off for field deployment.

5. Significance and Conclusion

This work establishes µTRec as a viable, non-intrusive safeguards tool for next-generation microreactors.

• Feasibility: It proves that missing fuel and internal configuration changes can be detected in sealed cores using realistic cosmic-ray fluxes within practical acquisition times.
• Efficiency: By leveraging physics-informed trajectory estimation and momentum data, the method reduces the required muon count (and thus imaging time) by orders of magnitude compared to conventional geometric methods.
• Deployment Readiness: The algorithm's robustness to detector resolution and the finding that a single spectrometer suffices for high performance make it a practical candidate for remote, autonomous verification of microreactor integrity.

The study provides a clear pathway toward deploying muon scattering tomography as a standard verification tool for the emerging class of portable, sealed nuclear energy systems.