Data Science and Machine Learning Technique for Predicting Electrical Resistivity in Recycled Concrete with Nopal as Addition

Abstract

Nowadays, the solid residues of concrete are considered waste, and this have been transformed into an environmental problem. This study analyzes the use of aggregates that comes from the concrete demolition process in order to create recycled concrete. The use of this material reduces costs and mitigates pollution. The present research describes the comparison of concrete blends using Opuntia Ficus Indica as an addition and recycled coarse aggregates as a substitution against a control blend. Mechanical and non-destructive tests were performed to evaluate the performance of each mixture. A data science technique was used to generate artificial data to increase the number of data to be evaluated. Numerical models were established to find correlations between all the features that describe the materials. The use of recycled aggregates and the use of the Opuntia Ficus Indica improved the performance in all the tests made to the concrete. Additionally, different models based on regression trees were used to predict with high accuracy the compressive strength in this kind of material just considering the electrical resistivity as an input parameter.