Incorporating non-destructive UPV into machine learning models for predicting compressive strength in SCM concrete

Abstract

The utilization of blast furnace slag (BFS) as supplementary cementitious material (SCM) has garnered significant attention due to its potential to reduce ordinary Portland cement consumption and associated CO2 emissions. This study focuses on predicting the compressive strength of SCM concrete containing BFS using various machine learning (ML) models. Specifically, artificial neural network (ANN), Support Vector Machine (SVM), and Decision Tree (DT) models were constructed, with blast furnace slag to binder ratio (BFS/B), age (A), and ultrasonic pulse velocity (UPV) as input parameters, and compressive strength as the output. Prior to model development, extensive statistical analysis was conducted on the dataset. Hyperparameter optimization via grid search was employed to enhance model performance. Evaluation metrics were utilized to compare the predictive capabilities of the models. Results revealed that the ANN model outperformed other ML models in predicting the compressive strength of SCM concrete containing BFS.