Structural Health Monitoring in Smart Structures Through Time Series Analysis
MetadataShow full item record
This paper describes the application of a structural health monitoring technique based on electrical measurements obtained by piezoceramics (PZT) patches bonded in lightweight structures. The goal is to detect and locate imminent structural change occurrence with statistical confidence through a nondestructive evaluation test. Though the major focus in damage detection is given by monitoring electrical impedance in frequency-domain, the current research work applies a novel approach based on time-series. In such case, auto-regressive moving average with exogenous input (ARMAX) system identification models and statistical process control (SPC) charts are used for linear prediction to detect and locate damages. In order to compare the results, the classical damage metric chart obtained by frequency response from input-output data is described. The efficacy of the proposed approach is demonstrated through experimental tests.