Accurately predicting the degree of hydration in cement systems blended with slag and fly ash is difficult when experimental data are limited. Empirical kinetic models typically need recalibration for each specific blend, while purely data-driven approaches often fail to maintain physical consistency. In this work, we present a physics-informed neural network for predicting isothermal hydration by incorporating the Avrami–Erofeev–Arrhenius ordinary differential equation into the training loss and using a composition-focused sub-network to relate blend proportions to blend-specific kinetic parameters. The model was trained and tested on 29,379 calorimetry data points from 77 multi-component cement systems gathered from two open-access datasets. For previously unseen test systems, it reached an $R^2$ of 0.9864 and a symmetric mean absolute percentage error (sMAPE) of 25.3%. Its overall sMAPE was lower than that of a neural network with the same architecture but without physics-based constraints (34.9%); stage-resolved analysis showed that the largest difference occurred during early hydration. Across eight random seeds, the physics constraint did not confer a data-efficiency or training-stability advantage over the same-architecture network. The distinguishing features of the PINN are instead its explicit ODE and monotonicity regularization and its composition-conditioned effective parameters. Although random forest regression produced lower pointwise error, it failed to maintain physically consistent hydration-rate behavior. Overall, the framework provides a composition-conditioned surrogate that incorporates kinetic regularization within the composition range investigated.

Equivalent thermal calculations describe mean heat transfer, but their accuracy may not extend to local thermal-bridge responses in 3D-printed ribbed walls. We compared the ISO 6946 path-layer method with two-dimensional finite-volume solutions for 2700 dry, steady, insulation-filled periodic cells. Four geometry and conductivity parameters were varied at fixed wall thickness, matrix conductivity, and surface resistances. Numerical benchmarks, mesh checks and ANSYS APDL cross-checks of five selected cases (resistance differences within 0.14%) assessed implementation. Experimental validation was not performed. Resistance error was within 5% for 72.2% of the full grid but for only 6.7% of a 75-case subset selected using literature-reported geometric and conductivity ranges. These proportions characterize the selected domains, not constructed-wall populations. The conductivity contrast at which the resistance error crossed 5% varied with geometry. Resistance errors also changed with wall scale and surface resistance. Sensitivity calculations with air cores used prescribed equivalent resistances without resolving cavity convection or radiation; ISO cavity-geometry compliance was not confirmed. Within the full-grid 5% band, the peak-to-mean heat-flux ratio reached 2.72, and the equivalent model overpredicted minimum internal surface temperature by up to 3.21 K. Global resistance accuracy is therefore not a proxy for local thermal accuracy: conditional error maps can screen resistance, but minimum surface temperature and peak heat flux need explicit-field assessment.