Effective Latent Heat Participation in PCM Cementitious Walls: Split-Cycle Calibration, Cross-Study Validation, and Post-Construction Follow-Up

Translating calorimetric latent heat into reliable thermal inputs for phase change material (PCM) cementitious building envelopes requires evaluation under finite thermal cycles. This study examines an effective latent-enthalpy multiplier using three published datasets for model identification, external comparison, and post-construction follow-up. Graph-reconstructed temperatures were analyzed using a one-dimensional enthalpy-conservative model, with heating-only calibration and withheld-cooling evaluation. Sensitivity analyses, parameter profiles, resampling, and benchmarks assessed robustness. Heating identified η ≈ 0.198, corresponding to a composite latent term of 3.99 kJ·kg−1 against the nominal 20.13 kJ·kg−1. With η fixed, cooling root-mean-square error (RMSE) was 0.337 °C, compared with 0.580 and 1.368 °C for no-latent and full-nominal models. In a separate hot-box case, heating-only reference calibration yielded raw post-shutoff RMSEs of 1.558 and 1.579 °C for transferred η and no latent heat; resampling did not establish a robust advantage. Reference-assisted correction improved agreement but remained conditional reconstruction. Six complete days approximately ten years after construction showed mean surface and indoor peak reductions of 6.65 and 2.25 °C, without establishing retained latent capacity. The results support cycle-specific calibration and cooling holdout testing for building-envelope modeling, while identifying η as a model-conditioned inverse descriptor rather than a measured utilization fraction or universal material property.