Proposed multi-scale models to predict the compressive strength of ultra-high performance fiber-reinforced concrete

Ultra-high performance fiber-reinforced concrete (UHPFRC) represents a new class of steel fiber-reinforced concrete with exceptional mechanical properties. It is ideal for construction projects that emphasize safety and protection. Based on its key components, developing an accurate and reliable model to predict the compressive strength (sigma c) of UHPFRC is cost-effective and time-efficient. Such models also provide valuable insights into optimal curing conditions and the appropriate schedule for formwork removal. In this study, three predictive models were developed: the linear relationship (LR) model, the nonlinear relationship (NLR) model, and the multi-logistic regression (MLR) model. A dataset of 312 samples was analyzed to examine the influence of 12 key variables on sigma c, including water-to-cement ratio (w/c), cement content (C), sand content (S), superplasticizer (SP), silica fume (SF), curing time (t), curing temperature (T), fiber content (Fb), fiber aspect ratio (AR), fiber diameter (Df), and fiber length (Lf). The models were evaluated using several statistical metrics, including the coefficient of determination (R2), root mean squared error (RMSE), mean absolute error (MAE), scatter index (SI), and objective (OBJ) value. Among the models, NLR-2 outperformed the others, achieving OBJ and SI values that were 24.85% and 19.62% lower, respectively, compared to the LR model. Additionally, sand content, fiber content, and curing time were identified as the most significant factors affecting the compressive strength of UHPFRC mixtures.