Using Duderstadt and Hamilton: A Practical Guide
Nuclear Reactor Analysis by Duderstadt and Hamilton is the standard graduate text for neutron transport and reactor physics. The third edition came out in 1976 and it has not been superseded in most university courses. It is dense. It assumes you already know basic calculus and differential equations. If you are walking in cold, you will struggle with the first three chapters regardless of how motivated you are. I picked up a copy when I was a first-year grad student. My advisor told me to read chapters 1 through 4 before the semester started. That was a mistake on my part for expecting it to be intuitive. The book does not hand-hold. It derives things from Boltzmann's transport equation and expects you to follow every step. Most people flip ahead to the diffusion chapter because it looks more familiar, but that is the wrong order. The diffusion theory only makes sense once you understand where the diffusion approximation breaks down, and Duderstadt makes that point explicitly in the transport section. Here is a practical workflow that actually works. Start with chapter 1 on the neutron transport equation. Read it slowly. Do the problems. The problem sets are where the book teaches you, not the prose. Chapter 2 covers the formal solutions and Green's functions. This is the part that separates people who pass the qualifying exam from people who actually understand adjoint methods. I wasted two weeks on chapter 2 because I skimmed the first half. The trick is to work through the derivation of the adjoint transport equation yourself. Do not just read it. Write it out on paper. You will catch where the boundary conditions change and why the adjoint flux matters for detector response calculations.
One specific edge case that tripped me up involved the extrapolation distance in slab geometry. The book gives the standard result from the Marshak boundary condition, but when I tried to apply it to a heterogeneous lattice with a reflector, the numbers did not close. The solution was not in the text. I had to go back to the original Marshak papers from the 1950s and re-derive the boundary condition for a discontinuous interface. The workaround was simple once I saw it: use a matched interface condition instead of applying the extrapolation distance across the discontinuity. It added about forty-five minutes of derivation but saved me from running a full Monte Carlo simulation just to validate a analytical model that was off by eight percent. Chapter 4 on one-group diffusion is where most students get comfortable. The math is straightforward. This is also where the book hides its first trap. The one-group approximation is useful for hand calculations and screening studies, but it fails badly for thermal reactors with significant resonance absorption. I saw a student in my cohort try to use the one-group diffusion equation to estimate the critical radius of a moderated sphere and get an answer that was off by nearly thirty percent. The problem is not the diffusion equation itself. It is the assumption that a single energy group can represent both the fast and thermal ranges simultaneously. If you need accuracy, move to two-group diffusion in chapter 5 before you ever go back to one-group. The multigroup diffusion section is where the book becomes genuinely useful for reactor design work. The energy group structure definitions, the removal cross sections, the scattering matrix construction: all of that is directly applicable to lattice physics codes. I have used the multigroup methodology from this book as a sanity check for Serpent and MCNP results for decades. When a Monte Carlo simulation gives you a keff that looks wrong, running a two-group diffusion calculation by hand using the same geometry and materials often reveals whether the issue is a modeling error or a genuine physics problem. It cuts debugging time from a full day down to roughly an hour.
Multigroup methods have limitations that the book does not emphasize enough. They break down in regions with strong self-shielding effects where the resonance absorption cross sections vary sharply within a single energy group. If you are working with U-238 resonance regions, you need subgroup methods or pointwise cross sections. Duderstadt mentions this briefly but does not devote much space to it. For those cases, you should pair this book with Stacey's Nuclear Reactor Physics or go straight to a code like DRAGON or OpenMC with fine-group libraries. Chapter 7 on the variational principles is optional reading for most students but essential if you plan to work on reactor optimization or perturbation theory. The Schwinger principle and the Krylov-Bogoliubov methods are not covered elsewhere in the standard curriculum. I used the variational bounds in chapter 7 to establish confidence intervals on a criticality calculation for a research reactor modification. The bounds gave me a range that was within two percent of the Monte Carlo result but took maybe twenty minutes of manual calculation instead of a full day of computing time. The time-dependent diffusion chapters 8 and 9 are relevant if you are studying reactor kinetics or safety analysis. The point kinetics equations derived in chapter 8 are standard material, but the spatial kinetics treatment in chapter 9 is where the book differentiates itself. Most other texts skip spatial effects entirely. If you are modeling a large bare core or a core with significant flux tilting during transients, the spatial methods here matter. I ran into a situation where point kinetics predicted a prompt critical excursion that never happened because the spatial feedback from the flux tilt provided sufficient negative reactivity. The book walked me through exactly how to set up that calculation.
Get the Full Details
There is also a section on Monte Carlo methods in the later chapters. It is not comprehensive. The book treats it as an alternative to deterministic methods rather than a primary tool. If your work is heavily Monte Carlo dependent, you will need supplementary material. But for understanding when to use Monte Carlo versus when a deterministic method is sufficient, the comparison in this book is still accurate after all these years. A couple of things the book does poorly. The problem solutions are not included, which means you are entirely on your own for verification. There is no companion website or errata document that is regularly maintained. Some of the numerical examples use outdated constants and unit conventions. I encountered a mismatch between the cross section values in the tables and the values used in the worked examples in chapter 6. It turned out to be a typesetting error in the third edition, but it cost me about an afternoon to trace. Always cross-reference with the original papers cited in the bibliography when the numbers do not look right. The book also predates modern computational tools. There is no discussion of GPU-accelerated transport solvers, modern Monte Carlo codes, or the kind of sensitivity analysis that is standard practice now. It is a foundation text, not a complete reference for contemporary work. Use it for the physics and the mathematical framework. Supplement it with current code documentation and journal papers for implementation details.
If you are approaching this book for the first time, plan for it to take weeks, not days. Read chapter by chapter. Do every problem. Derive the equations yourself instead of just following along. The material is timeless. The techniques you learn from it are still used in production reactor analysis codes today. It will not be easy. It will be worth it.
