KeffNet: Neural Neutron Transport — Monte Carlo Simulation Data

This dataset contains simulation data and results for the paper "KeffNet: Neural Neutron Transport (KNNT) — A Physics-Informed Neural Network for Solving the Neutron Diffusion Eigenvalue Problem." Files: - ground_truth_fission_rates.csv: Ground truth fission rates for pure isotopes - predictions_final_debug.csv: Model predictions for diverse nuclear materials - analytic_keff_test.csv: Analytic keff values for validation - collocation_points.csv: Collocation points used for training - geometry_regions.csv: Geometry definitions for fuel and reflector regions - interface_points.csv: Points at the fuel-reflector interface - materials.csv: Material cross-section data - validation_flux.csv: Flux validation data All data was generated through Monte Carlo simulations and PINN training. Results are fully reproducible using the code available at: https://github.com/drgon19940-afk/monte-carlo-neutron-ml本数据集包含论文《KeffNet：神经中子输运（Neural Neutron Transport，KNNT）——用于求解中子扩散本征值问题的物理感知神经网络》的模拟数据与计算结果。 文件列表： - ground_truth_fission_rates.csv：纯同位素的真实裂变率数据 - predictions_final_debug.csv：面向多样化核材料的模型预测结果 - analytic_keff_test.csv：用于验证的解析有效增殖因子数据 - collocation_points.csv：训练过程中使用的配点数据 - geometry_regions.csv：燃料区与反射层的几何定义数据 - interface_points.csv：燃料-反射层界面处的点位数据 - materials.csv：材料截面参数数据 - validation_flux.csv：中子通量验证数据 所有数据均通过蒙特卡洛模拟与物理感知神经网络（Physics-Informed Neural Network，PINN）训练生成。 本数据集的全部结果均可通过公开于https://github.com/drgon19940-afk/monte-carlo-neutron-ml 的代码实现完全复现。