Prediction of flexural strength of natural pozzolana and limestone blended concrete using machine learning based models

Abstract

Natural pozzolana and limestone ternary blended concrete is a new type of limestone calcined clay cement (LC3) concrete. Natural pozzolana and limestone can benefit each other and contribute to the performance improvement of concrete. Flexural strength is an important engineering property of concrete. This study predicts the development of flexural strength of ternary blended concrete using machine learning based models, such as artificial neural networks (ANN) and gene expression programming (GEP). Concrete mixtures and curing ages are used as input parameters of ANN and GEP. Flexural strength is output parameter of ANN and GEP. The correlation coefficients of GEP and ANN model are 0.98 and 0.99, respectively. GEP and ANN can make reliable evaluations of flexural strengths of ternary blended concrete.