This study investigates the electrical resistivity behavior of cementitious composites incorporating carbon fiber (CF), carbon black (CB), and their hybrid combinations under varying curing ages, humidity levels (0–100%), and temperatures (0°C to 120°C). 16 different mixtures are prepared with CF contents up to 0.9 vol.% and CB contents up to 9 wt.% of cement. Experimental results reveal that hydration-induced densification significantly increases resistivity in plain cement paste (from 267.5 to 999.5 Ω·cm over 28 days), whereas hybrid CF–CB composites maintain low and stable resistivity values (3.9–7.1 Ω·cm), demonstrating superior environmental robustness. Moisture loss and sub-zero temperatures markedly increase resistivity in the control specimen, while conductive fillers preserve electrical continuity. At elevated temperatures, hybrid composites exhibit a thermally stable conductive response, in contrast to the signal degradation observed in the control due to thermal cracking. Machine learning models (XGBoost, SVR, and MLP) are employed to predict resistivity based on five input variables, achieving high predictive accuracy, with XGBoost reaching an R 2 of 0.981 on test data. SHAP analysis identifies carbon black as the dominant contributor to conductivity, quantitatively validating the synergistic CF–CB mechanism. These findings demonstrate a scalable pathway for designing environmentally resilient, self-sensing cementitious materials supported by data-driven modeling.
Tabansiz-Goc et al. (Mon,) studied this question.