1. School of Computer Applications, Guilin University of Technology, Guilin 541004, China
2. Centrale Mediterranee, Chateau-Gombert Technology Park, Marseille 13451, France
3. Guangxi Key Laboratory of Digital Infrastructure, Guangxi Zhuang Autonomous Region Information Center, Nanning 530201, China
4. Joint International Research Laboratory of Spatio-temporal Information and Intelligent Location Services, Guilin University of Electronic Technology, Guilin 541004, China
lilong@guet.edu.cn
Show less
History+
Received
Accepted
Published Online
2026-02-11
2026-06-11
2026-11-18
PDF
(25162KB)
i
This article has been updated.A Correction to this article was published on .
Accurate prediction of ultra-high-performance concrete (UHPC) compressive strength with quantified uncertainty is essential for reliability-based structural design, yet existing machine learning approaches predominantly provide deterministic point estimates without calibrated confidence intervals. This study develops a comprehensive Gaussian process (GP) regression framework that integrates native uncertainty quantification with model interpretability for UHPC strength prediction. Eleven GP configurations were systematically evaluated against seven conventional machine learning algorithms using 863 experimental observations encompassing diverse fiber types and mixture parameters. The GP with a radial basis function kernel using automatic relevance determination achieved the best performance among the GP models (coefficient of determination (R2) = 0.9147, root mean square error = 10.16 MPa), with a 95% prediction interval coverage of 95.4%, closely matching the nominal level. Comparative analysis revealed that Bootstrap uncertainty quantification applied to categorical boosting achieved only 60.7% coverage at the 95% confidence level, demonstrating severe miscalibration unsuitable for engineering reliability assessment. Statistical robustness was validated through 30 repeated experiments, with Wilcoxon signed-rank tests confirming significant differences (p < 10−8) between GP and gradient boosting methods. GP-based Shapley additive explanations (SHAP) and GP-based accumulated local effects methods were developed to propagate prediction uncertainty through interpretability calculations, identifying curing age as the dominant predictor (SHAP: (15.15 ± 4.40) MPa) with strong cross-method correlation (ρ = 0.896). The framework demonstrated noise robustness (3.1% R2 degradation under 20% label corruption) and data efficiency (95% prediction interval coverage probability above 86% across all training set sizes). Engineering applications including exceedance probability computation, P (compressive strength > design threshold), enable direct reliability assessment, while active learning analysis revealed that the top 20% high-uncertainty samples capture 29% of total uncertainty, providing cost-effective experimental validation strategies. The probabilistic framework addresses the critical gap between point prediction accuracy and decision-oriented risk assessment in UHPC mixture optimization.
Kravanja G , Mumtaz A R , Kravanja S . A comprehensive review of the advances, manufacturing, properties, innovations, environmental impact and applications of ultra-high-performance concrete (UHPC). Buildings, 2024, 14(2): 382
[2]
Wang S W , Wang B , Zhu H T , Chen G , Li Z Z , Yang L , Zhang Y K , Zhou X M . Ultra-high performance concrete: Mix design, raw materials and curing regimes—A review. Materials Today Communications, 2023, 35: 105468
[3]
Liu J , Wei J , Li J , Su Y , Wu C Q . A comprehensive review of ultra-high performance concrete (UHPC) behaviour under blast loads. Cement and Concrete Composites, 2024, 148: 105449
[4]
Niu F Y , Liu Y H , Xue F C , Sun H , Liu T , He H J , Kong X G , Chen Y T , Liao H J . Ultra-high performance concrete: A review of its material properties and usage in shield tunnel segment. Case Studies in Construction Materials, 2025, 22: e04194
[5]
Yoo D Y , Banthia N , Yoon Y S . Recent development of innovative steel fibers for ultra-high-performance concrete (UHPC): A critical review. Cement and Concrete Composites, 2024, 145: 105359
[6]
Du W H , Yu F , Qiu L S , Guo Y X , Wang J L , Han B G . Effect of steel fibers on tensile properties of ultra-high-performance concrete: A review. Materials, 2024, 17(5): 1108
[7]
Li J Y , Deng Z C . Tensile behavior of ultra-high performance concrete reinforced with different hybrid fibers. Structural Concrete, 2023, 24(1): 1415–1435
[8]
Muhyaddin G F . Mechanical and fracture characteristics of ultra-high performance concretes reinforced with hybridization of steel and glass fibers. Heliyon, 2023, 9(7): e17926
[9]
Yin T Y , Liu K N , Fan D Q , Yu R . Derivation and verification of multilevel particle packing model for Ultra-High Performance Concrete (UHPC): Modelling and experiments. Cement and Concrete Composites, 2023, 136: 104889
[10]
Yin T Y , Yu R , Liu K N , Wang Z Y , Fan D Q , Wang S Y , Feng Y , Shui Z H . Precise mix-design of Ultra-High Performance Concrete (UHPC) based on physicochemical packing method: From the perspective of cement hydration. Construction and Building Materials, 2022, 352: 128944
[11]
Li W X , Zhao Y , Zhang Y S , Xie Z C , Zhang J F , Huang F G , Meng L K , He Z Y , Xia J L , Zhang Y . et al. Optimized mix design method of ultra-high performance concrete (UHPC) and effect of high steel fiber content: Mechanical performance and shrinkage properties. Journal of Building Engineering, 2024, 97: 110746
[12]
Yu R , Zhou F J , Yin T Y , Wang Z Y , Ding M X , Liu Z , Leng Y , Gao X , Shui Z H . Uncovering the approach to develop ultra-high performance concrete (UHPC) with dense meso-structure based on rheological point of view: Experiments and modeling. Construction and Building Materials, 2021, 271: 121500
[13]
Xi J Y , Liu J Z , Yang K , Zhang S H , Han F Y , Sha J F , Zheng X . Role of silica fume on hydration and strength development of ultra-high performance concrete. Construction and Building Materials, 2022, 338: 127600
[14]
Xu D S , Tang J H , Hu X , Zhou Y C , Yu C , Han F Y , Liu J P . Influence of silica fume and thermal curing on long-term hydration, microstructure and compressive strength of ultra-high performance concrete (UHPC). Construction and Building Materials, 2023, 395: 132370
[15]
Neamat S , Haido J H , Wardeh G . Hydration and micro-structure of UHPC containing different pozzolanic materials: A review. AIP Conference Proceedings, 2024, 2944(1): 020007
[16]
Li J Q , Wu Z M , Shi C J , Yuan Q , Zhang Z H . Durability of ultra-high performance concrete—A review. Construction and Building Materials, 2020, 255: 119296
[17]
Saladi N , Montanari L , De La Varga I , Spragg R , Graybeal B . Assessing durability properties of ultra-high performance concrete-class materials. Materials and Structures, 2023, 56(8): 155
[18]
Fan D Q , Zhu J Y , Fan M X , Lu J X , Chu S H , Dong E L , Yu R . Intelligent design and manufacturing of ultra-high performance concrete (UHPC)—A review. Construction and Building Materials, 2023, 385: 131495
[19]
Huu Nguyen M , Nguyen T A , Ly H B . Ensemble XGBoost schemes for improved compressive strength prediction of UHPC. Structures, 2023, 57: 105062
[20]
Das P , Kashem A . Hybrid machine learning approach to prediction of the compressive and flexural strengths of UHPC and parametric analysis with shapley additive explanations. Case Studies in Construction Materials, 2024, 20: e02723
[21]
Kumar R , Rai B , Samui P . A comparative study of prediction of compressive strength of ultra-high performance concrete using soft computing technique. Structural Concrete, 2023, 24(4): 5538–5555
[22]
Rasmussen C E . Gaussian processes in machine learning. Bousquet O, Luxburg U, Rätsch G, eds, 2004, 63–71
[23]
Qin F P , Chen Y P , Hou M D , Huang J B . Interpretable machine learning for flexural strength prediction of 3D-printed concrete incorporating supplementary cementitious materials. Buildings, 2026, 16(16): 3151
[24]
Mahjoubi S , Barhemat R , Meng W N , Bao Y . AI-guided auto-discovery of low-carbon cost-effective ultra-high performance concrete (UHPC). Resources, Conservation and Recycling, 2023, 189: 106741
[25]
Wakjira T G , Kutty A A , Alam M S . A novel framework for developing environmentally sustainable and cost-effective ultra-high-performance concrete (UHPC) using advanced machine learning and multi-objective optimization techniques. Construction and Building Materials, 2024, 416: 135114
[26]
Abellán-García J . Study of nonlinear relationships between dosage mixture design and the compressive strength of UHPC. Case Studies in Construction Materials, 2022, 17: e01228
[27]
Abuodeh O R , Abdalla J A , Hawileh R A . Assessment of compressive strength of Ultra-high Performance Concrete using deep machine learning techniques. Applied Soft Computing, 2020, 95: 106552
[28]
Marani A , Jamali A , Nehdi M L . Predicting ultra-high-performance concrete compressive strength using tabular generative adversarial networks. Materials, 2020, 13(21): 4757
[29]
Khan M , Lao J C , Dai J G . Comparative study of advanced computational techniques for estimating the compressive strength of UHPC. Journal of Asian Concrete Federation, 2022, 8(1): 51–68
[30]
Nguyen N H , Abellán-García J , Lee S , Garcia-Castano E , Vo T P . Efficient estimating compressive strength of ultra-high performance concrete using XGBoost model. Journal of Building Engineering, 2022, 52: 104302
[31]
Ferdosian I , Camões A . Eco-efficient ultra-high performance concrete development by means of response surface methodology. Cement and Concrete Composites, 2017, 84: 146–156
[32]
Tabani A , Biswas R . Assessment of compressive strength of ultra-high-performance concrete using advanced machine learning models. Structural Concrete, 2025, 26(6): 7269–7283
[33]
Farouk A I B , Zhu J S . Prediction of interface bond strength between ultra-high-performance concrete (UHPC) and normal strength concrete (NSC) using a machine learning approach. Arabian Journal for Science and Engineering, 2022, 47(4): 5337–5363
[34]
Rasmussen C E , Williams C K I . Gaussian Processes for Machine Learning. Cambridge: The MIT Press, 2006,
[35]
Schulz E , Speekenbrink M , Krause A . A tutorial on Gaussian process regression: Modelling, exploring, and exploiting functions. Journal of Mathematical Psychology, 2018, 85: 1–16
[36]
Balachandran P V , Xue D Z , Theiler J , Hogden J , Lookman T . Adaptive strategies for materials design using uncertainties. Scientific Reports, 2016, 6(1): 19660
[37]
Chen N Z , Guedes Soares C . Reliability assessment for ultimate longitudinal strength of ship hulls in composite materials. Probabilistic Engineering Mechanics, 2007, 22(4): 330–342
[38]
Belfekih T , Fitas R , Schaffrath H J , Schabel S . Graph-based analysis for the characterization of corrugated board compression. Materials, 2024, 17(24): 6083
[39]
Efron B , Tibshirani R J . An Introduction to the Bootstrap. New York: Chapman and Hall, 1994,
[40]
Mooney C Z, Duval R D. Bootstrapping. Thousand Oaks: SAGE Publications, 1993
[41]
Palmer D S , O’Boyle N M , Glen R C , Mitchell J B O . Random forest models to predict aqueous solubility. Journal of Chemical Information and Modeling, 2007, 47(1): 150–158
[42]
Tuo R, Wang W J. Uncertainty quantification for Bayesian optimization. In: Proceedings of the 25th International Conference on Artificial Intelligence and Statistics. Valencia: PMLR, 2022, 2862–2884
[43]
Jiang Y , Hui Y , Wang Y , Peng L L , Huang G Q , Liu S Y . A novel eigenvalue-based iterative simulation method for multi-dimensional homogeneous non-Gaussian stochastic vector fields. Structural Safety, 2023, 100: 102290
[44]
Lundberg S M, Lee S I. A unified approach to interpreting model predictions. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc., 2017, 4768–4777
[45]
Apley D W , Zhu J Y . Visualizing the effects of predictor variables in black box supervised learning models. Journal of the Royal Statistical Society Series B: Statistical Methodology, 2020, 82(4): 1059–1086
[46]
Foulser-Piggott R . Quantifying the epistemic uncertainty in ground motion models and prediction. Soil Dynamics and Earthquake Engineering, 2014, 65: 256–268
[47]
La H , Nguyen T . Interpretable and optimized TabNet-TPE for predicting bond strength in corroded reinforced concrete using a global database. Structures, 2025, 80: 109707
[48]
Bolbolvand M , Tavakkoli S M , Alaee F J . Prediction of compressive and flexural strengths of ultra-high-performance concrete (UHPC) using machine learning for various fiber types. Construction and Building Materials, 2025, 493: 143135
[49]
Luat N V , La H , Nguyen T . Deep ensemble-based probabilistic modeling of torsional capacity in reinforced concrete beams with uncertainty decomposition. Applied Mathematical Modelling, 2026, 157: 116856
[50]
La H , Luat N V , Lee K , Nguyen T . Uncertainty-aware prediction of shear strength in reinforced concrete deep beams using quantile machine learning and explainable artificial intelligence. Structures, 2026, 84: 110907
[51]
La H , Le-Thanh T , Nguyen T . Reliability-informed inverse design of dual tunnels with deep evidential regression. Reliability Engineering & System Safety, 2026, 269: 112134
[52]
La H , Uddin M , Nguyen T . Reliable inverse optimization of rice husk ash concrete via copula-based design spaces and conformal quantile surrogates. Expert Systems with Applications, 2026, 301: 130582