PINN-BC Reproducibility Package for Temperature–Frequency Equivalent-Resistance Baseline Calibration of an Inductive Oil-Debris Sensor

This dataset and reproducibility package supports the study "PINN-Inspired Calibration of the Temperature–Frequency Equivalent-Resistance Baseline of an Inductive Oil-Debris Sensor." The package contains raw experimental measurement records, processed datasets, analysis and model-training scripts, model evaluation results, frozen reference outputs, independent reproduction results, and figure-generation materials. The study develops PINN-BC, a physics-informed neural network-inspired baseline calibration model that estimates the instrument-reported equivalent-resistance baseline from temperature and excitation frequency. The evaluation includes 36 operating conditions from Config C, six-fold leave-one-temperature-out validation, three random seeds, classical regression baselines, ablation experiments, and robustness analyses. The materials are provided to support transparency, reproducibility, and independent verification of the reported results.