3D Electroquasistatic Core–Shell Dosimetry and Global LHS Uncertainty Quantification Pipeline v2.0

Overview This repository contains the complete computational reproducibility package for the manuscript: "A Reproducible Electroquasistatic Computational Framework for Three-Dimensional Internal Dosimetry: Numerical Verification and Global Uncertainty Quantification in a Core–Shell Tissue Phantom," submitted to Computer Methods and Programs in Biomedicine. The archive provides a conditional three-dimensional computational dosimetry framework. It evaluates electric-field and current-density redistribution in an idealized, PDAC-inspired core-shell phantom, utilizing a frequency-domain electroquasistatic (EQS) solver. The methodology includes a three-layer concentric-sphere analytical verification benchmark (yielding 0.45% relative L2 error on a 256³ Cartesian mesh), mesh refinement convergence testing, and a global screening analysis using Latin Hypercube Sampling (LHS) and nonparametric bootstrapping (N=10,000). Archive Contents This deposit (v2.0) replaces the earlier analytical prototype with the full 3D Cartesian finite-difference solver. It includes the standalone Python suite used to generate the numerical datasets, conduct the statistical analysis (Spearman rank correlation coefficients), and render the final publication-ready figures, alongside their corresponding high-resolution outputs: complete_3d_eqs_solver,_verification_&_figure_suite_(zenodo_v2_0).py – Master Python script containing the 3D ILU-BiCGStab EQS solver, analytical Legendre verification, LHS pipeline, and automated figure generation. mesh_convergence.csv – Output dataset tracking L2 error and convergence rates across 64³, 128³, and 256³ grids. lhs_10000_samples.csv – Raw dataset of the 10,000 Latin Hypercube evaluations. uq_bootstrap_summary.csv – Statistical summary of the LHS pipeline, including median attenuation (-11.6 dB) and Spearman correlations. Figure_1 (KB).pdf through Figure_4 (KB).pdf – High-resolution vector graphics of the bounded domain, mesh convergence, 3D field redistribution, and global uncertainty distributions. Usage & Dependencies The Python script can be executed as a standalone file to automatically run the analytical limit checks, solve the 3D finite-difference system, process the LHS surrogate, and generate Figures 1 through 4. The code requires standard scientific Python libraries: numpy, scipy, and matplotlib. Physics Model Scope & Disclaimers This repository evaluates a mathematical model mapping dosimetric sensitivity and field redistribution in abstract bounding geometries under explicit low-frequency assumptions. It does not claim measured low-frequency PDAC properties, assert a clinical treatment prediction, define exposure protocols, validate an in vivo shielding mechanism, or assert the safety or feasibility of any specific biological or pharmacological intervention.