Optimal Accerlerometer Locations for Structural Health Monitoring using Time-Domain System Identification

Soon Jung Kwon, Hae Sung Lee, Soo Bong Shin

ABSTRACT

 This paper presents two algorithms for determining optimal accerlerometer locations for structural health monitoring when structural condition is assessed by a system identification scheme in time-domain. The accerlerometer locations are determined by ranking the components of an effective independent distribution vector computed from a Fisher information matrix. One of the proposed algorithms formulates a Fisher information matrix by multiplying acceleration matrix with its transpose and the other as a Gauss-Newton hessian matrix composed of acceleration sensitivities with respect to structural parameters. Since the structural parameters cannot be known exactly in an actual application, a statistical approach is proposed by setting an error bound between the actual and the baseline values. To examine the algorithm, simulation studies have been carried out on a two-span planer truss. The results using locations selected by the two algorithms were compared.

[Prev] [List] [Next]

If you need full text, please visit http://www.scientific.net