Modal identification technique based on distributed sensor networks

Abstract

Identification of the dynamic characteristics of civil structures is important in structural health monitoring. A large number of data must be available from a dense array of sensors for large-scale civil structures and poses a big challenge to the conventional centralized processing technique. Smart sensor networks (SSN) with decentralized processing capability provides new possibilities for structural health monitoring. A distributed method is proposed to calculate the global mode shape in SSN. Stochastic subspace identification is implemented to identify local mode shapes, which are rescaled by using particle swarm optimization method, and subsequently to combine global mode shapes. Using an arch bridge model as an example, the distributed method is shown to be effective. The global mode shapes are close to those from centralized method according to modal assurance criterion (MAC).