Data-driven fresh property prediction and mix design optimization of self-compacting geopolymer concrete

Abstract

Self-compacting geopolymer concrete (SCGC) requires coordinated control of workability and strength during mix design, yet predicting fresh-state properties from mix proportions alone has not been reliably achieved. This study trained machine learning models on 327 two-part SCGC mixes from 37 published sources to predict five fresh properties (slump flow, T500, V-funnel time, L-box ratio, J-ring step) and 28 d compressive strength (CS28d). Curing temperature, pre-demolding duration, and post-curing regime were added as inputs for CS28d, giving 15 inputs total. Five tree-based algorithms (random forest, gradient boosting, extra trees, XGBoost, LightGBM) were compared, with hyperparameters tuned via RandomizedSearchCV or Optuna. Under random-split cross-validation, extra trees achieved CV R2 of 0.949 (V-funnel), 0.910 (L-box), and 0.934 (J-ring); gradient boosting led for slump flow at CV R2 = 0.929. Under leave-one-source-out validation; which withholds entire laboratories from training; R2 fell to 0.261 for slump flow and 0.120 for CS28d; T500 reached R2 = − 0.279. The resulting gap, ΔR2 = 0.67–0.92 across outputs, measures the inter-laboratory information leakage that random-split validation conceals and that prior SCGC ML studies have not corrected for. Adding the curing inputs raised CS28d test R2 by 0.119. SHAP and permutation importance analysis identified curing temperature as the dominant driver of CS28d and produced physically consistent rankings for fresh property outputs. The trained models served as surrogates in a differential evolution framework that simultaneously optimizes for EFNARC workability classes (SF2/SF3, PA2) and minimum compressive strength targets, tested for ambient (25 °C, 24 h) and oven (70 °C, 48 h) curing across three strength thresholds. Eleven of twelve scenarios returned feasible solutions, with total binder content from 362 to 574 kg m−3.