Utilisation of Machine Learning Techniques to Model Creep Behaviour of Low-Carbon Concretes - Machine learning creep compliance prediction

Aug 5, 2026

Upload by Finn

Low-carbon concrete mixes using high volumes of fly ash and slag as cement replacements are increasingly adopted to decarbonise construction, yet they exhibit creep behaviour that differs from conventional concrete. Since creep strongly affects long-term structural response, serviceability and durability, a supervised machine learning framework is developed to predict creep compliance for concretes with fly ash and slag. Gaussian process regression, ANN, random forest regression and decision tree regression are trained on a dataset compiled from an existing comprehensive creep collection and literature, then evaluated with holdout validation (70% train, 30% validation). Statistical assessment shows random forest and Gaussian process regression deliver the best accuracy; sensitivity analyses indicate limited variability across training repetitions and input selections, Shapley analysis identifies time as most influential, and comparisons with experiments confirm accurate creep representation by GPR/RFR/ANN while DTR underperforms.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages25

Reading Time1 h 3 min

AccessFree to download

Introduction

Machine Learning Methods

Model Development and Training Data

Validation Strategy

Results and Model Performance

Accuracy Comparison

Sensitivity and Input Importance

Experimental Comparison

Conclusions

Utilisation of Machine Learning Techniques to Model Creep Behaviour of Low-Carbon Concretes This is the Published version of the following publication Bouras, Yanni and Li, Le (2023) Utilisation of Machine Learning Techniques to Model Creep Behaviour of Low-Carbon Concretes. Buildings, 13 (9) . ISSN 2075-5309 The publisher’s official version can be found at [https://www.mdpi.com/2075-5309/13/9/2252](https://www.mdpi.com/2075-5309/13/9/2252)[ ](https://www.mdpi.com/2075-5309/13/9/2252)Note that access to this version may require subscription. Downloaded from VU Research Repository [https://vuir.vu.edu.au/47204/](https://vuir.vu.edu.au/47204/) buildings Article Utilisation of Machine Learning Techniques to Model Creep Behaviour of Low-Carbon Concretes Yanni Bouras * and Le Li Citation: Bouras, Y.; Li, L. Utilisation of Machine Learning Techniques to Model Creep Behaviour of Low-Carbon Concretes. Buildings 2023, 13, 2252. [https://doi.org/](https://doi.org/)[ ](https://doi.org/)[10.3390/buildings13092252](10.3390/buildings13092252) Academic Editor: Jan Foˇrt Received: 7 August 2023 Revised: 28 August 2023 Accepted: 4 September 2023 Published: 5 September 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . Institute of Sustainable Industries and Liveable Cities, College of Sport, Health and Engineering, Victoria University, Melbourne 3011, Australia; [le.li@vu.edu.au](le.li@vu.edu.au) * [Correspondence: yanni.bouras@vu.edu.au](Correspondence: yanni.bouras@vu.edu.au) Abstract: Low-carbon concrete mixes that incorporate high volumes of ﬂy ash and slag as cement replacements are becoming increasingly more common as part of efforts to decarbonise the construction industry. Though environmental beneﬁts are offered, concretes containing supplementary cementitious materials exhibit different creep behaviour when compared to conventional concrete. Creep can signiﬁcantly impact long-term structural behaviour and inﬂuence the overall serviceability and durability of concrete structures. This paper develops a creep compliance prediction model using supervised machine learning techniques for concretes containing ﬂy ash and slag as cement substitutes. Gaussian process regression (GPR), artiﬁcial neural networks (ANN), random forest regression (RFR) and decision tree regression (DTR) models were all considered. The dataset for model training was developed by mining relevant data from the Infrastructure Technology Institute of Northwestern University's comprehensive creep dataset in addition to extracting data from the literature. Holdout validation was adopted with the data partitioned into training (70%) and validation (30%) sets. Based on statistical indicators, all machine learning models can accurately model creep compliance with the RFR and GPR found to be the best-performing models. The sensitivity of the GPR model's performance to training repetitions, input variable selection and validation methodology was assessed, with the results indicating small variability. The importance of the selected input variables was analysed using the Shapley additive explanation. It was found that time was the most signiﬁcant parameter, with loading age, compressive strength, elastic modulus, volume-to-surface ratio and relative humidity also showing high importance. Fly ash and silica fume content featured the least inﬂuence on creep prediction. Furthermore, the predictions of the trained models were compared to experimental data, which showed that the GPR, RFR and ANN models can accurately reﬂect creep behaviour and that the DTR model does not give accurate predictions. Keywords: artiﬁcial neural networks;

FAQs

Why is creep behaviour important for low-carbon concretes?

Creep can significantly affect long-term structural behaviour and influence serviceability and durability.

Which machine learning models were evaluated for creep compliance prediction?

Gaussian process regression (GPR), artificial neural networks (ANN), random forest regression (RFR) and decision tree regression (DTR) were all considered.

What factors most influence the Gaussian process regression model?

Shapley additive explanation results indicate time is the most significant parameter, with loading age, compressive strength, elastic modulus, volume-to-surface ratio and relative humidity also highly important.

Aug 5, 2026

Upload by Finn