Data-Driven Framework for The Flow Characteristics of Self-Compacting Concrete Using Machine Learning

Data-Driven Framework for The Flow Characteristics of Self-Compacting Concrete Using Machine Learning

ES评分 0

DOI 10.20900/jsr20260061
刊名
JSR
年,卷(期) 2026, 8(3)
作者
作者单位

Dipartimento di Ingegneria Informatica, Modellistica, Elettronica e Sistemistica-DIMES, University of Calabria, Via Pietro Bucci, Rende 87036, Italy Facultad de Ciencias Pecuarias, Escuela Superior Politécnica de Chimborazo (ESPOCH), Panamericana Sur km. 1 ½, Riobamba 060155, Ecuador Facultad de Ciencias, Escuela Superior Politécnica de Chimborazo (ESPOCH), Panamericana Sur km. 1 ½, Riobamba 060155, Ecuador Facultad de Mecánica, Escuela Superior Politécnica de Chimborazo (ESPOCH), Panamericana Sur Km 1 ½, Riobamba 060155, Ecuador Grupo de Investigación y Desarrollo de Nanotecnología, Materiales y Manufactura (GIDENM), Escuela Superior Politécnica de Chimborazo (ESPOCH), Panamericana Sur Km 1 ½, Riobamba 060155, Ecuador

摘要
This study used hybrid metaheuristic machine learning techniques to predict the workability characteristics of self-compacting concrete (SCC) mixes, focusing on the influence of coarse aggregate size. Three advanced techniques were applied: genetic programming via evolutionary approach optimization (GP-EAO), evolutionary polynomial regression via genetic algorithm optimization (EPR-GAO), and artificial neural networks via modified biological neuron optimization (ANN-MBNO). Because conventional concrete constituents and the maximum nominal size (MNS) of coarse aggregates critically influence both the fresh and hardened properties of concrete, they were integrated into the SCC database. The results demonstrated that MNS significantly impacts SCC workability, with smaller sizes (10–18.5 mm) enhancing metrics such as slump flow time, height, and diameter. Among the models, ANN-MBNO prediction demonstrated the highest accuracy, with R2, mean absolute error, and root mean squared error values of 0.964, 7 mm, and 8 mm, respectively, significantly outperforming GP-EAO and EPR-GAO. Furthermore, the ANN-MBNO model effectively captured the relative importance of all input parameters, emphasizing the critical role of MNS. This study provides a robust framework for predicting SCC workability, offering practical tools for sustainable production and optimized resource utilization.
Abstract
This study used hybrid metaheuristic machine learning techniques to predict the workability characteristics of self-compacting concrete (SCC) mixes, focusing on the influence of coarse aggregate size. Three advanced techniques were applied: genetic programming via evolutionary approach optimization (GP-EAO), evolutionary polynomial regression via genetic algorithm optimization (EPR-GAO), and artificial neural networks via modified biological neuron optimization (ANN-MBNO). Because conventional concrete constituents and the maximum nominal size (MNS) of coarse aggregates critically influence both the fresh and hardened properties of concrete, they were integrated into the SCC database. The results demonstrated that MNS significantly impacts SCC workability, with smaller sizes (10–18.5 mm) enhancing metrics such as slump flow time, height, and diameter. Among the models, ANN-MBNO prediction demonstrated the highest accuracy, with R2, mean absolute error, and root mean squared error values of 0.964, 7 mm, and 8 mm, respectively, significantly outperforming GP-EAO and EPR-GAO. Furthermore, the ANN-MBNO model effectively captured the relative importance of all input parameters, emphasizing the critical role of MNS. This study provides a robust framework for predicting SCC workability, offering practical tools for sustainable production and optimized resource utilization.
关键词
self-compacting concrete (SCC); workability/slump; coarse aggregate size; ANN-MBNO; GP-EAO; EPR-GAO
KeyWord
self-compacting concrete (SCC); workability/slump; coarse aggregate size; ANN-MBNO; GP-EAO; EPR-GAO
基金项目
页码 -
  • 参考文献
  • 相关文献
  • 引用本文

Fredy Romero, Kerly Vaca, María Albuja, Nestor Ulloa, Pablo Mancheno. Data-Driven Framework for The Flow Characteristics of Self-Compacting Concrete Using Machine Learning [J]. Journal of Sustainability Research. 2026; 8; (3). - .

  • 文献评论

相关学者

相关机构