Equivalent structural parameters based non-destructive prediction of sustainable concrete strength using machine learning models via piezo sensor

Abstract

As concrete is one of the most common material used in the construction industry, it is essential to monitor and predict the strength development during curing/hydration process in order to avoid unexpected catastrophic failure during the construction process. Hence, this paper presents equivalent structural parameters-based strength monitoring and prediction of ternary blended concrete system using machine learning (ML). Different piezo configurations were adopted to check their sensitivity and suitability in real-life applications and ML models were developed based on the extracted impedance data acquired using piezo sensors. Comparing the sensitivity of different piezo configurations, embedded configuration performed the best during the hydration process and strength gain. Furthermore, fine gaussian support vector machine (SVM) model best predicted the compressive strength with an error of less than 2% and coefficient of determination (R2) value of 1 and 0.99 for ternary blended and conventional concrete system, respectively. © 2021 Elsevier Ltd