POWER-LET: Physics-Informed Optimal Transport for Cross-Domain Radiation Risk Modelling Between Terrestrial Electromagnetic Fields and Space Radiation
The biological effects of radiation depend on the underlying mechanism of energy transfer. Ionising radiation produces direct molecular damage, whereas the biological effects of extremely low-frequency electromagnetic fields (ELF-EMF) remain uncertain and are generally associated with indirect cellular responses. This study presents a computational framework for comparing the biological effects of terrestrial ELF-EMF and ionising space radiation. A literature-informed correlation matrix was developed to compare four radiation environments: electrical substation ELF-EMF, Galactic Cosmic Rays (GCR), Solar Particle Events (SPE), and trapped radiation belts. Four biological endpoints were considered, namely DNA double-strand breaks, oxidative stress, membrane permeability damage, and chromosomal aberrations. The comparison shows consistently stronger associations between ionising radiation and biological damage than between ELF-EMF and the same endpoints. An Advanced Radiation Simulator Suite was developed to quantify radiation effects in both terrestrial and space environments. The terrestrial model incorporates electromagnetic field attenuation and exposure modelling. The space radiation model combines radiation shielding with established radiobiological dose-response models. The study also introduces the POWER-LET (Power-network Optimal Wasserstein Embedding for Radiation-Linear Energy Transfer) framework. The framework constructs a shared latent space between terrestrial EMF and space radiation descriptors using Optimal Transport. Physics-informed constraints are incorporated to ensure consistency with established physical laws. Uncertainty quantification is included to evaluate the reliability of cross-domain predictions. The simulation results demonstrate a clear separation between ionising and non-ionising radiation environments while preserving physically meaningful relationships within the latent representation. The uncertainty analysis provides a transparent assessment of prediction confidence for cross-domain mapping. The proposed framework offers an integrated approach for comparative radiation assessment and provides a foundation for future studies in radiation biology, space medicine, radiation protection, and electrical power systems.
Authors
- Amos S. Kubeka (ORCID: https://orcid.org/0000-0002-9185-7549)
- King Kumire
Institutions
- University of South Africa (ZA)
Publication Details
- Journal
- American Journal of Applied Scientific Research
- Published
- 2026-11-14
- DOI
- https://doi.org/10.11648/j.ajasr.20261204.11
- Primary Topic
- Electromagnetic Fields and Biological Effects
- Type
- article
- Field-Weighted Citation Impact
- 0.00