ABSTRACT
A 1-D CNN regression model is designed for autonomously monitoring stress in concrete specimens utilizing the raw impedance signatures of smart aggregate (SA) sensors. The fundamental theory of the impedance measurement model of SA is presented. Next, the compression experiment on SA-embedded concrete cylinders is carried out, and the impedance signals of the cylinders are recorded under different stress levels. The 1-D CNN regression model learned the impedance signals for predicting the concrete stress is constructed. Then, the average performance of the proposed model is verified via the 10-fold cross-validation method. Consequently, the feasibility of the developed model is investigated under the effect of noises in signals and reduction of the input training data.
