A Low-Cost Sensing Solution for SHM, Exploiting a Dedicated Approach for Signal Recognition

• Published in: Sensors (Basel, Switzerland) Vol. 24; no. 12; p. 4023
• Main Authors: Andò, Bruno, Greco, Danilo, Navarra, Giacomo, Lo Iacono, Francesco
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 01.06.2024; MDPI
• Subjects: Algorithms; Behavior; Classification; Data transmission; Economic aspects; embedded architecture; Embedded systems; Emergency communications systems; Investigations; Medical care, Cost of; Methods; Nondestructive testing; Preventive maintenance; Public buildings; sensing system; Sensors; Signal processing; signals classification; structural health monitoring; time–frequency analysis; User interface; Wavelet transforms; time–frequency analysis; signals classification; structural health monitoring; sensing system; embedded architecture
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s24124023

Health assessment and preventive maintenance of structures are mandatory to predict injuries and to schedule required interventions, especially in seismic areas. Structural health monitoring aims to provide a robust and effective approach to obtaining valuable information on structural conditions of buildings and civil infrastructures, in conjunction with methodologies for the identification and, sometimes, localization of potential risks. In this paper a low-cost solution for structural health monitoring is proposed, exploiting a customized embedded system for the acquisition and storing of measurement signals. Experimental surveys for the assessment of the sensing node have also been performed. The obtained results confirmed the expected performances, especially in terms of resolution in acceleration and tilt measurement, which are 0.55 mg and 0.020°, respectively. Moreover, we used a dedicated algorithm for the classification of recorded signals in the following three classes: noise floor (being mainly related to intrinsic noise of the sensing system), exogenous sources (not correlated to the dynamic behavior of the structure), and structural responses (the response of the structure to external stimuli, such as seismic events, artificially forced and/or environmental solicitations). The latter is of main interest for the investigation of structures’ health, while other signals need to be recognized and filtered out. The algorithm, which has been tested against real data, demonstrates relevant features in performing the above-mentioned classification task.