Assessing the compressive strength of self-compacting concrete with recycled aggregates from mix ratio using machine learning approach

Abstract

The requirement of the construction sector pushes researchers and academicians to determine the 28-day concrete compressive strength due to less consumption of natural products and reduced cost. One recommended method to reduce the cost and simultaneously adopt sustainability is introducing recycled aggregates in concrete. Most typical structures require concrete, which is self-flowable and compactable; specific structures require concrete, which is self-flowable and compactable; one such concrete is self-compacting concrete (SCC). 515 mix design samples for SCC with recycled aggregates are collected from the literature and used for training, validation, and testing to create the model using machine learning techniques (Extra Gradient (XG) Boosting, Ada Boosting, Gradient Boosting, Light Gradient Boosting, Category Boosting, K Nearest Neighbors, Extra Trees, Decision Trees, Random Forest, and Support Vector Machine). The correlation between input and output variables is analyzed using ANOVA and is indicated that data can be used to develop machine learning models successfully. Sensitive analysis and error assessment are performed to choose the best machine learning methods, and it found that CB, KNN, and ERT have the highest R2 value and lowest MSE.

Keywords

• Compressive strength
• Recycled aggregate
• Self-compacting concrete
• Machine learning
• Correlation
• ANOVA