Development of an Optimization Algorithm for Designing Low-Carbon Concrete Materials Standardization with Blockchain Technology and Ensemble Machine Learning Methods

Guardado en:

Detalles Bibliográficos
Publicado en:	Buildings vol. 15, no. 16 (2025), p. 2809-2830
Autor principal:	Ebenezer, Nwetlawung Zilefac
Otros Autores:	Yi-Hsin, Lin
Publicado:	MDPI AG
Materias:	Green development Carbon content Accuracy Datasets Collaboration Deep learning Performance prediction Algorithms Artificial neural networks Blockchain Construction contracts Machine learning Carbon Learning algorithms Viscosity Artificial intelligence Concrete curing Root-mean-square errors Genetic algorithms Neural networks Sustainability Variables Construction industry Cement Federated learning Design optimization Optimization algorithms Compressive strength
Acceso en línea:	Citation/Abstract Full Text + Graphics Full Text - PDF
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Descripción
Resumen:	This study presents SmartMix Web3, a framework combining ensemble machine learning and blockchain technology to optimize low-carbon concrete design. It addresses two key challenges: (1) the limitations of conventional models in predicting concrete performance, and (2) ensuring data reliability and overcoming collaboration issues in AI-driven sustainable construction. Validated with 61 real-world experiments in Cameroon and 752 mix designs, the framework shows major improvements in predictive accuracy and decentralized trust. To address the first research question, a stacked ensemble model comprising Extreme Gradient Boosting (XGBoost)–Random Forest and a Convolutional Neural Network (CNN) was developed, achieving a 22% reduction in Root Mean Square Error (RMSE) for compressive strength prediction and embodied carbon estimation compared to traditional methods. The 29% reduction in Mean Absolute Error (MAE) results confirms the superiority of Extreme Learning Machine (EML) in low-carbon concrete performance prediction. For the second research question, SmartMix Web3 employs blockchain to ensure tamper-proof traceability and promote collaboration. Deployed on Ethereum, it automates verification of tokenized Environmental Product Declarations via smart contracts, reducing disputes and preserving data integrity. Federated learning supports decentralized training across nine batching plants, with Secure Hash Algorithm (SHA)-256 checks ensuring privacy. Field implementation in Cameroon yielded annual cost savings of FCFA 24.3 million and a 99.87 kgCO2/m3 reduction per mix design. By uniting EML precision with blockchain transparency, SmartMix Web3 offers practical and scalable benefits for sustainable construction in developing economies.
ISSN:	2075-5309
DOI:	10.3390/buildings15162809
Fuente:	Engineering Database