Application of SHM Using an Autonomous Sensor Network

Abstract

Assessment of the condition of a structure in operation and subsequently its remaining maintenance free service time, is of increasing importance for various fields of engineering. The main drivers for this are cost effectiveness, increased system reliability, system safety and reduced environmental impact. The current generation of monitoring systems relies on active, power intensive excitation and wired communication. Systems based on operational vibrations of the structure and employing a network of smart, autonomously operating and wireless sensors offer new possibilities, but also pose new constraints. Damage identification methods are therefore sought that combine local low power usage and low data transmission with a high reliability. The focus of this paper is on operational vibrations and modal based Structural Health Monitoring damage identification methods, applied in large civil structures such as wind turbine towers and bridge decks and, to a lower extent, in large composite structures. Three methods are compared, both experimentally and numerically: Peak Picking (PP), Random Decrement – Frequency Domain Decomposition (RD–FDD) and Random Decrement – covariance based Stochastic Subspace Identification (RD–SSIcov). The RD–FDD method is found to be a suitable method for modal based damage identification, given the restrictions on smart wireless sensor networks.