Xuefei Tang | Searcharxiv

arXiv2026

RAYTHEIA: A high-performance ray-tracing algorithm for three-dimensional direction-dependent equations in astronomical simulations

We present RAYTHEIA, a high-performance reverse ray-tracing algorithm designed to efficiently solve three-dimensional direction-dependent equations in astronomical simulations. The algorithm uses a dual-grid framework in which the native simulation mesh -- serving as the source grid for ray emission -- and an adaptive mesh refinement (AMR) Cartesian contribution grid are constructed for efficient ray-walking and contribution accumulation. The core of the algorithm integrates a leaf-only linear-octree data structure to reduce memory overhead, the digital differential analyzer (DDA) traversal method to efficiently determine the ray-walking path, Morton Code indexing to fast leaf cell lookup during traversal, and the slab method to analytically compute the path length. Furthermore, RAYTHEIA employs a hybrid (MPI/OpenMP) distributed parallel framework with a chunk-to-chunk communication strategy, achieving exceptional, near-ideal linear speed-up ratio and delivering high-end performance. We integrate RAYTHEIA with the 3D-PDR code to solve the complex chemistry and radiation transfer in photodissociation regions (PDRs). This allowed the modelling of three-dimensional PDR chemistry in a turbulent, star-forming cloud at an unprecedented resolution of $512^3$ grid cells. The algorithm demonstrates accuracy and convergence even at low angular resolutions. We further showcase the capabilities of RAYTHEIA by producing high-resolution synthetic emission maps of key diagnostic lines of a star-forming region capturing physical effects such as [O I] $63\mu$m self-absorption, measuring the [C I]-bright but CO-dark molecular gas, and deriving a CO-to-H$_2$ conversion factor in agreement with observations.

astro-ph.IM↗

arXiv2025

NeuralPDR: Neural Differential Equations as surrogate models for Photodissociation Regions

Computational astrochemical models are essential for helping us interpret and understand the observations of different astrophysical environments. In the age of high-resolution telescopes such as JWST and ALMA, the substructure of many objects can be resolved, raising the need for astrochemical modeling at these smaller scales, meaning that the simulations of these objects need to include both the physics and chemistry to accurately model the observations. The computational cost of the simulations coupling both the three-dimensional hydrodynamics and chemistry is enormous, creating an opportunity for surrogate models that can effectively substitute the chemical solver. In this work we present surrogate models that can replace the original chemical code, namely Latent Augmented Neural Ordinary Differential Equations. We train these surrogate architectures on three datasets of increasing physical complexity, with the last dataset derived directly from a three-dimensional simulation of a molecular cloud using a Photodissociation Region (PDR) code, 3D-PDR. We show that these surrogate models can provide speedup and reproduce the original observable column density maps of the dataset. This enables the rapid inference of the chemistry (on the GPU), allowing for the faster statistical inference of observations or increasing the resolution in hydrodynamical simulations of astrophysical environments.

astro-ph.GA↗