Machine Learning Techniques for Low-Carbon Concrete Strength Prediction
SCOPUS 2025 = 2.4

Indexed: WEB of SCIENCE,
SCOPUS, DOAJ, SCIELO,…
PDF

How to Cite

Jing, M., & Sun, P. (2026). Machine Learning Techniques for Low-Carbon Concrete Strength Prediction. Latin American Journal of Solids and Structures, 23(8), e9056. https://doi.org/10.1590/1679-7825/e9056

Abstract

The strength of Low-Carbon Concrete (LCC) is affected by various factors like the mix composition, curing environment, and carbon emissions (CE), which leads to difficulties in making precise predictions. This paper introduces a Machine Learning (ML) Low-Carbon Concrete Strength Prediction (LCCSP) technique via a Self-Organizing MapRadial Basis Function Neural Network (SOM-RBFNN). Construction materials (CMs) are selected plus age and water to cement ratio (W/C) as the three main variables for the model inputs based on CE modeling and performance analysis. Firstly, the SOM network groups and finds feature in the input data that are multi-dimensional; then the cluster centers and radii obtained are used as the hidden-layer parameters of the RBFNN, which enhances its efficiency in recognizing highly nonlinear relationships. Test results show that the proposed model is able to maintain a low Mean Absolute Error (MAE) and a Coefficient of Determination (R) close to 1 in both single-factor and multi-factor scenarios, which implies a strong capability to accurately predict, persistently and be flexible to changes in the LCC strength.

PDF
Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright (c) 2026 Ming JING, Peng SUN

Downloads

Download data is not yet available.