Pulsed Height Tally in mixed neutron-gamma fields

[kosbor-personal]

Intro

Hi everyone!

Long time listener, first time caller. This is my first Pull Request, so I would appreciate any feedback on how I can improve this contribution and also potential future ones.

All the best,
Bor Kos

Problem Description

We’ve been working on verifying and validating OpenMC for well-logging applications. In this pull request, I wanted to address an issue we encountered when we started analysing so‑called pulsed‑neutron applications, where we are interested in gamma spectra from capture reactions from 14 MeV neutrons thermalized within a rock formation.

Long story short, these have been our observations regarding pulse height tallies (PHT) in gamma-only and coupled neutron–gamma fields:

• Correct behaviour for gamma‑only PHT cases (i.e. perfect match with MCNP and experimental results)
• Excellent agreement between OpenMC and MCNP for gamma spectra in coupled neutron–gamma simulations
• Significant deviations in pulse height tallies for coupled neutron–gamma simulations

This issue was shared on the forum:
https://openmc.discourse.group/t/pulse-height-tally-in-mixed-neutron-gamma-fields/6229

This led me to believe that there is an issue in how neutron‑generated photons are treated within the pulse height accounting logic.

Looking at the code, I believe photons produced by neutrons were incorrectly having their energy subtracted from pht_storage, leading to systematic undercounting in PHT results.

In pht_secondary_particles(), OpenMC subtracts energy from pht_storage for all secondary photons, assuming their parent’s energy had previously been added to the PHT. This assumption is valid for photon/electron/positron parents but invalid for neutron reactions, because:

• Neutron energy is never added to pht_storage
• Neutron‑generated photons therefore had energy subtracted without ever being added

This results in a net loss of recorded pulse height.

Fix

The fix ensures that energy is only subtracted for secondary particles whose parent particle is either a photon, electron, or positron. This is achieved by tracking the parent particle type for each secondary particle and only applying the energy subtraction when the parent is a photon/electron/positron.

Code Changes

1. include/openmc/particle_data.h

   • Added parent_particle to SourceSite
   • Added corresponding member variable and accessors to ParticleData

2. src/particle.cpp

   • Set parent_particle in create_secondary()
   • Propagate via from_source()
   • Apply whitelist logic in pht_secondary_particles()

3. openmc/lib/core.py

   • Added parent_particle to the Python _SourceSite ctypes structure

Testing / Validation

• Pytest suite

  • Failed tests matched failures in unmodified openmc-dev.
  • Suspected cause is compiler settings.
  • None of the failed tests are related to pulse height tallies.

• Gamma‑only PHT

  • Results remain unchanged.
  • Plots attached for two different sources:
    • Potassium (left)
    • Thorium (right)

• Coupled neutron–gamma simulations
  • Pulsed‑neutron tools tested in two environments:
    • Dolomite
    • Brine
  • Results match MCNP.
  • Lesser effect observed in brine, most likely due to high chlorine absorption.

[GuySten]

I think you can add a guard in this line (skip it for neutrons) instead of adding another bank attribute.
This should significantly simplify this PR.