Real time damage detection through single low-cost smart sensor

Continuous vibration-based structural monitoring is increasingly used to assess the state of health of existing structures and infrastructures in environmental conditions through non-inva- sive methods that allow, in most cases, an early identification of damage. To date, the innova- tive wireless smart sensor networks are the subject of numerous researches in the field of structural health monitoring, but their use on real structures is still limited, due to problems related to energy consumption and algorithmic optimization. In fact, most of the traditional identification algorithms work in centralized topology and are not suitable for electronic ele- ments with low computational capacity. Nevertheless, the high energy consumption of wireless communication does not allow continuous data transmission for centralized processing in real time. However, the limited costs of new technologies, compared to traditional wired acquisition systems, shifts the interest towards innovative solutions, both from an algorithmic and hard- ware point of view, in order to provide innovative monitoring systems that can also be used on minor structures, for which traditional systems would be inaccessible. This paper presents a first practical application of a structural identification algorithm specifically designed for low- cost embedded electronic systems on a scaled laboratory model. The proposed solution consists of a single sensing node able to identify natural frequencies in real time, even in conditions of non-stationary excitation and variable structural characteristics. The estimated parameters are periodically uploaded to a cloud platform, where a preliminary real-time damage detection takes place.