Predicting early‐age stress evolution in restrained concrete by thermo‐chemo‐mechanical model and active ensemble learning

Early-age stress (EAS) is an important index for evaluating the early-age cracking risk of concrete. This paper encompasses a thermo-chemo-mechanical (TCM) model and active ensemble learning (AEL) predicting EAS evolution. The TCM provides data AEL model. First, based on Fourier's law, Arrhenius’ equation, rate-type creep built to simulate heat transfer, cement hydration, viscoelasticity, which together determine Then, material composed eXtreme Gradient Boosting adjusted Model Code 2010 allow parametric study database construction. Finally, framework built, incorporates principal component analysis (PCA), Gaussian process, light gradient boosting machine (LGBM). resulted in following findings: (1) dimensionality 672-by-1 vector can be effectively reduced by PCA, first (PC) global representing magnitude EAS; (2) mechanical field validated testing data. Correlation PC quantifies influence various input parameters model, accordance with common understandings evolution process. (3) one-shot (OSEL) both achieve high prediction performance set, whose R2 reaches 0.961 0.948, respectively. Thanks uncertainty-based query procedure, comparing OSEL, shows advantages over whole training history. (4) significantly reduce number samples required training, major improvement efficiency considering computational cost