Towards a cloud-based platform for structural health monitoring: implementation and numerical issues

Structural Health Monitoring (SHM) is increasingly important in protecting and maintaining architectural heritage. Its main goal is to distinguish ordinary fluctuations in a building's response from other, possibly anomalous, behaviour. SHM starts setting sensors to measure accelerations or velocities and other environmental parameters over time at fixed points of the structure. The time series processing makes it possible to perform modal tracking and damage/anomaly detection while correlating dynamical and environmental parameters. In practice, these activities are conducted separately, using different numerical codes. Thus, the idea is to take the first step to distance from such practice, leveraging the MOSCARDO system, which encompasses a Wireless Sensor Network (WSN) and a platform designed according to a cloud architecture that provides services for storing and processing data from the WSN. We employ a code based on the Stochastic Subspace Identification (SSI) technique to improve the system's capabilities, and we exploit the SSI's theoretical features to get an efficient implementation that will be integrated into the cloud-based platform. This pipeline is here presented considering data collected from a monitoring campaign on the "Matilde donjon" in Livorno (Italy) and reporting preliminary numerical results on the identification of the modal parameters.