A combined technique of implantable sensors and probabilistic localization method for monitoring acoustic events on concrete slab

Abstract

Continuous monitoring of acoustic events in concrete such as cracking and impacting, and providing their location information is a significant topic in structural health monitoring (SHM). This paper develops a combined technique of concrete implantable cube (CIC) and probabilistic localization method to monitor acoustic sources on concrete slabs. The CIC that integrates an octahedron-shaped piezoelectric sensor array can be implanted in concrete structures in long term and supply an omni-dimensional sensing capacity. Aiming at the designed sensor array and the impacts of randomness and heterogeneity of concrete material on the localization accuracy, a probabilistic localization method was proposed. To demonstrate the feasibility of the proposed technique, an experiment including 48 impact tests was conducted on a concrete slab specimen. The acoustic signals received by the implanted CIC were used to reconstruct localization maps to predict the impact areas. The results were quantified using three matrixes including accuracy, precision, and recall; consequently, the accuracy obtained the highest average score close to 0.98 and performed most steadily. In addition, to optimize the proposed technique furtherly, the sensitivity of two key parameters associated with this method were discussed. Overall, the experimental study showed a great potential of this developed technique for long-term monitoring acoustic events on concrete structures.