Real-Time Structural Damage Identification of Time-Varying Systems

Abstract

This paper presents an investigation of real-time structural damage identification for time-varying systems, such as aircraft with changing flight condition. The on-line damage detection is implemented by comparing the changes of the identified model, which is updated by synthesizing time-domain input/output data, with the changes of the analytical model due to damage. The study is based on the identified models with various model structures, state space model, transfer function, and Auto-Regressive eXogenous (ARX) model. Some identified parameters are much more sensitive to noise, and the inclusion of these parameters in damage identification may significantly reduce the reliability of damage detection with false indications of damage. A technique is presented for the selection of the identified parameters for damage identification. The investigation is demonstrated by application to the Benchmark Active Controls Technology (BACT) wind-tunnel model with time-varying dynamic pressure and Mach number.