Oct 2026· Journal of materials in civil engineering· 0 citations· 83 references
Abstract
Ultrahigh-performance concrete (UHPC) offers superior mechanical properties and durability but is constrained by high density, cost, and environmental impact due to its cement-intensive composition. This study presents a machine learning (ML)–driven multiobjective optimization framework for UHPC mix design, integrating waste aggregates, supplementary cementitious materials (SCMs), and performance-enhancing components to balance strength and sustainability. A robust data set combining 694 literature-derived and 87 experimental data points underpins the framework. Unsupervised anomaly detection (isolation forest) is employed to refine data quality, while ML techniques identify key parameters influencing UHPC properties. Predictive models, including artificial neural networks, random forests, and light gradient-boosting machine (LightGBM), are trained on full and reduced feature sets to ensure accuracy and generalizability. LightGBM, showing the best predictive performance, is embedded into single-objective and multiobjective optimization processes. Single-objective optimization achieves rapid and accurate convergence for individual performance targets. Multiobjective optimization yields diverse Pareto-optimal solutions, exposing trade-offs among strength, cost, and embodied carbon. Experimental validation confirms that optimized mixes improve performance while reducing environmental footprint. Two novel UHPC mixes were experimentally validated: (1) an SCM-based formulation that reduces clinker use while maintaining structural integrity; and (2) a UHPC mix incorporating waste glass aggregates, enhancing circularity and lowering carbon emissions. This study provides a scalable, data-driven approach to ecoefficient UHPC design, enabling intelligent, application-ready solutions that align with construction performance demands and global sustainability goals.
The compressive strength of ultra-high-performance concrete (UHPC) is jointly influenced by multiple factors, including material composition, mixture proportion parameters, and curing regime. Conventional empirical methods are therefore insufficient to accurately characterize the highly nonlinear relationships involved. To improve the prediction accuracy of UHPC compressive strength and to achieve mixture proportion optimization that simultaneously considers mechanical performance, economic efficiency, and environmental impact, this study developed random forest (RF), artificial neural network (ANN), gradient boosting decision tree (GBDT), and extreme gradient boosting (XGBoost) models based on 810 publicly available UHPC experimental datasets. Model performance was evaluated using R2, RMSE, MAE, and MAPE. To enhance the robustness of model validation, repeated K-fold cross-validation, sensitivity analysis with different random seed splits, and benchmark model comparisons were further introduced. The results indicate that the XGBoost model achieved superior predictive performance on both the test set and robustness validation, with test-set R2, RMSE, MAE, and MAPE values of 0.9604, 7.77, 5.58, and 4.80, respectively. The model was further interpreted using SHAP, PDP, and ICE methods, and the results revealed that curing age, fiber content, silica fume content, and water-to-binder ratio were important variables affecting the compressive strength of UHPC. Furthermore, XGBoost was used as a surrogate model and coupled with NSGA-II and TOPSIS methods for multi-objective optimization. Under the constraints of compressive strength, water-to-binder ratio, superplasticizer-to-binder ratio, and absolute volume, a computationally recommended UHPC mixture proportion balancing strength, cost, and carbon emissions was obtained. This study provides a reproducible machine-learning-assisted approach for UHPC compressive strength prediction and low-carbon, cost-effective mixture proportion design.
Rong Li, Teng Zhou, Siyu Lu et al.· Applied Sciences· 0 citations
The novel compression-cast concrete (CCC) delivers superior mechanical and durability performance over conventional vibration-cast concrete (VCC), alongside economic and environmental advantages. However, its widespread adoption requires an optimized and systematic design method. This study presents a data-driven framework that integrates machine learning (ML) and multiobjective optimization for both forward prediction and inverse design of CCC. Using an experimental data set, various ML models were trained, with Optuna-optimized backpropagation neural networks (OP_BPNN) showing the best accuracy. Model interpretability was enhanced using individual conditional expectation and Shapley additive explanations. The validated OP_BPNN served as a surrogate in inverse optimization via nondominated sorting genetic algorithm III (NSGA-III), targeting compressive strength while minimizing cost and
CO
2
emissions and maximizing density. Optimal solutions were ranked using the technique for order of preference by similarity to ideal solution (TOPSIS). Compared to VCC, the optimized CCC showed up to 15% potential reduction in cost and 38% lower
CO
2
emissions, as predicted by the model within the studied parameter range. A user-friendly graphical interface was developed to facilitate practical implementation. The framework offers a scalable tool for CCC design aligned with project-specific performance and sustainability goals.
Muhammad Tahir, Ying-Wu Zhou, Biao Hu et al.· Journal of materials in civi...· 0 citations
Ultra-high-performance concrete (UHPC) offers exceptional mechanical and durability properties but often relies on quartz powder, raising sustainability and occupational health concerns. This study introduces an integrated experimental-computational framework for predicting the compressive strength of UHPC and developing quartz-free mixtures. Experimentally, the effects of mixing sequence, sand characteristics, superplasticizer chemistry, and curing regime were investigated, leading to a quartz-free UHPC achieving 136 MPa at 28 days under heat-curing. A dataset of 550 UHPC compressive strength records was compiled, incorporating quantitative mix proportions and categorical variables (cement type, superplasticizer base, fiber type, and specimen geometry). Sixty-three machine learning models from tree-based, boosting, and support vector machine families were optimized using seven meta-heuristic algorithms. The Particle Swarm Optimization-tuned XGBoost model achieved the highest prediction accuracy (R2 = 0.897, RMSE = 7.63 MPa), followed by the Differential Evolution-optimized Random Forest (R2 = 0.867, RMSE = 8.70 MPa). SHapley Additive exPlanations (SHAP) analysis identified curing age as the most influential predictor after optimization. The proposed framework enables accurate and interpretable UHPC strength prediction and supports the design of safer and more sustainable quartz-free UHPC with reduced experimental effort.
Mohamed Ayman, Amr Elnemr· Scientific Reports· 0 citations
Ultra-high performance concrete (UHPC), an advanced cementitious composite characterized by superior mechanical properties and durability, requires multi-objective collaborative optimization in mix proportion design to facilitate its large-scale engineering applications. To address this challenge, this study proposes an intelligent design framework for UHPC driven by the synergistic integration of Stacking ensemble learning and the non-dominated sorting genetic algorithm III (NSGA-III). In the model construction stage, a Stacking ensemble model was developed by integrating eight heterogeneous algorithms as base learners and employing linear regression as the meta-learner. Subsequently, six hyperparameter optimization strategies were employed to fine-tune the model. At the optimization decision stage, the optimal Stacking ensemble model was embedded into the NSGA-III algorithm, and the technique for order preference by similarity to ideal solution (TOPSIS) was employed to select the optimal mix design from the Pareto front. The results demonstrate that the predictive accuracy of the Stacking ensemble model significantly outperforms that of the individual single-algorithm models. Furthermore, multi-objective optimization conducted for three target strength grades, UC100, UC120, and UC140, revealed that all optimized schemes successfully meet the specified strength requirements while effectively balancing environmental impact and production cost. This research establishes a data-driven, multi-criteria intelligent decision-making pathway for the multi-objective design of UHPC, with substantial implications for advancing the interdisciplinary integration of materials science and artificial intelligence.
Wei Zhang, Chao Liu, Yi-Zhou Yao et al.· Journal of Building Design a...· 0 citations
Ultra-high-performance concrete (UHPC) exhibits exceptional mechanical properties and durability. However, its compressive strength is highly dependent on complex mix design parameters. While traditional experimental techniques and regression-based models are commonly used to evaluate UHPC compressive strength, machine learning approaches offer an efficient alternative for capturing complex nonlinear relationships. This study develops a machine learning–based framework to predict the compressive strength of UHPC and compares the predictive performance of five advanced algorithms: Extremely Randomized Trees (ER), Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost), CatBoost, and Artificial Neural Network (ANN). A comprehensive experimental database was utilized for training and validation purposes. Among the evaluated models, CatBoost achieved the best predictive performance, with a coefficient of determination (R²) exceeding 0.90, a root mean square error (RMSE) of approximately 4.5 MPa, and a mean absolute error (MAE) of approximately 3.6 MPa. However, subgroup residual analysis showed that the prediction reliability was not uniform across the full strength range. In particular, mixtures with compressive strength ≥180 MPa exhibited larger errors and systematic underprediction, mainly due to the limited number of ultra-high-strength samples in the compiled database. Therefore, the model is more reliable within well-represented strength ranges, while predictions in the ultra-high-strength region should be interpreted with caution. SHAP-based analysis, feature dependency analysis, and both Individual Conditional Expectation (ICE) and Partial Dependence Plots (PDP) were employed. These explainable AI techniques identified key variables and quantified their contributions to the compressive strength of UHPC. The findings demonstrate that interpretable machine learning can support preliminary UHPC mixture assessment by combining predictive performance with physically meaningful insights.
Nga T. T. Nguyen, T. Nguyen, Tuan-Khoi Nguyen et al.· PLoS ONE· 0 citations
Ultra-High-Performance Concrete (UHPC) is a state-of-the-art concrete technology with exceptional qualities, including high compressive strength (CS) and durability. The CS, an essential property of UHPC, is determined through costly, time-consuming studies that require large amounts of material. To overcome these constraints, this work aimed to estimate the CS of UHPC using a range of single- and hybrid-machine learning (ML) methods. For this, five ML models, including Decision Tree (DT), Gradient Boosting (GB), Light Gradient Boosting Machine (LightGBM), CatBoost, and a stacking Ensemble model combining these base models, were developed. The input space includes cement, silica fume, slag, fly ash, quartz powder, limestone powder, nano-silica, water, fine and coarse aggregate, fiber, superplasticizer, temperature, relative humidity, and age. The findings demonstrate that the Ensemble model achieved the best overall predictive performance across 20 Monte Carlo simulations, with a mean RMSE of 6.751 ± 0.316 MPa, MAE of 4.943 ± 0.219 MPa, and R² of 0.972 ± 0.002. According to the findings of 1D, 2D partial dependence plot (PDP), and SHAP analyses, age, cement, silica fume, water, sand, fiber, and superplasticizer were the primary variables influencing UHPC's CS.
May Huu Nguyen, Hai-Van Thi Mai, Ho-Ang Son· Journal of Science and Trans...· 0 citations