Public concrete datasets often contain duplicate records, coupled variables, and cross-source distribution shifts, which may lead to overly optimistic model evaluation. Based on a deduplicated UCI high-performance concrete dataset (1005 samples), this study develops a leakage-controlled data-driven workflow with applicability-domain assessment. TabPFN, SHAP, and NSGA-II are used for compressive strength prediction, model-response attribution, and scenario-based candidate mix screening, respectively. Model evaluation follows a unified data split, inner training-set cross-validation, and an independent test-set protocol. In addition, 502 non-overlapping records from the Mendeley PCC dataset are used as an external benchmark to examine cross-source transferability and sensitivity to distribution shift. The results show that TabPFN achieves the highest R² and the lowest RMSE, MAE, and MAPE on the internal UCI test set, with values of 0.953, 3.744 MPa, 2.265 MPa, and 7.580%, respectively; however, its advantage over strong baselines such as CatBoost is limited. On the external Mendeley PCC dataset, TabPFN remains competitive, with R², RMSE, and MAE values of 0.490, 15.175 MPa, and 11.457 MPa, respectively, but its performance is close to that of random forest, XGBoost, and CatBoost. The 5NN applicability-domain stratification shows that external samples located within the 95% 5NN applicability domain achieve improved performance (R² = 0.634 and RMSE = 12.367 MPa), suggesting that external prediction errors are associated with the distance from the source-domain distribution. SHAP results indicate that cement, ground granulated blast-furnace slag, curing age, and water are the main attribution variables in the model output; their response directions should be interpreted as statistical attributions rather than material causal mechanisms. The Pareto candidate mixes generated by NSGA-II satisfy basic engineering constraints. Nevertheless, because the external benchmark reveals sensitivity to cross-source distribution shift, the resulting mix proportions should be treated as pre-experimental screening candidates rather than engineering-validated low-GWP concrete mix proportions.

This study examines the effect of air-entraining agents (AEAs) type on cement-mortar air content and air-void structure under reduced atmospheric pressure. Six representative AEAs—cetyltrimethylammonium bromide (CTAB), triterpenoid saponin (TS), sodium dodecylbenzenesulfonate (SDBS), sodium abietate (SA), cocamidopropyl betaine (CAB), and fatty alcohol polyoxyethylene ether (AEO-9)—were selected. Their foaming ability and time-dependent foam stability were measured in deionized water and in cement filtrate, and the air content of fresh mortars and the distribution of air-voids in hardened mortars were determined at 100 and 60 kPa. The results show that, at 100 kPa, TS, CAB, and CTAB produced higher initial foam height and better foam stability in deionized water than AEO-9, SA, and SDBS. TS and CAB also maintained a higher number density of bubbles and slower coalescence. In addition, all surfactant systems showed lower initial foam height and stability in cement filtrate than in deionized water, with SDBS, SA, and AEO-9 experiencing the greatest declines. When the pressure decreased from 100 kPa to 60 kPa, the mortar air content dropped by 8–15%, with the smallest reduction for TS (~8%) and the largest for CTAB (~15%). At 60 kPa, air voids with radius < 250 μm decreased markedly in hardened mortars: by 51%, 25%, and 28% for the control, CTAB, and AEO-9 mortars, respectively; but only by 14% for TS, highlighting its superior retention of fine air voids. Overall, amphoteric/saponin-type systems (represented by TS) exhibit better tolerance and stabilization, and are recommended for high-altitude concrete.