Deep learning based automatic crack detection for concrete structures using piezoelectric smart aggregates

Abstract

Identification of cracks in concrete structures is a significant study area of evaluating the concrete health over the last 20 years. Various ingredients, including binders, fly ash, silica fume, pulverized granulated slag from blast furnaces, mica fume, and silica, might influence the SCC’s burn resistance. The presentation finishes with an experimental demonstration of the practical application of the Piezoelectric Smart Aggregates (PSA) sensor. When using sensors to identify and monitor cracks, prior knowledge of fracture locations and angles is not required, according to the results. The proposed PSA model has better accuracy R > 0.8 and fewer error values RMSE and MAE. The silica fume concrete attains comparatively larger tensile strengths of 2.33 and compressive strength of 63.7 MPa than the other concrete mix. Overall, the proposed method ensures reliability, structural durability, and an efficient alternative to the existing models for evaluating early crack detection. Additionally, the study enhances the ceramic composites optimization for improved performance, emphasizing the admixtures in enhancing the mechanical strength and conductivity in next-generation concrete structures.