Singular spectrum analysis for identifying structural nonlinearity using free-decay responses : Application detection and diagnosis in composite laminates

Garcia, David and Trendafilova, Irina (2014) Singular spectrum analysis for identifying structural nonlinearity using free-decay responses : Application detection and diagnosis in composite laminates. In: 26th International Conference on Noise and Vibration Engineering, 2014-09-15 - 2014-09-17.

[thumbnail of Garcia D & Trendafilova I - Pure - Singular spectrum analysis for idenifying structural nonlinearity using free-decay responses - Application fo delamination detection and diagnosis in composite laminates Sep 2014] PDF. Filename: Garcia_D_Trendafilova_I_Pure_Singular_spectrum_analysis_for_idenifying_structural_nonlinearity_using_free_decay_responses_Application_fo_delamination_detection_and_diagnosis_in_composite_laminates_Sep_2014.pdf
Accepted Author Manuscript
Download (355kB)

Abstract

This investigation considers a methodology for analysis of the vibratory response of composite laminate structures which is based on Singular Spectrum Analysis. Composite laminate structures generally demonstrate nonlinear dynamic behaviour as a result of their intrinsic material nonlinear nature. Since the nonlinearities on a number of occasions induce relativity small changes in the vibratory response which are difficult to identify, the raw measured dynamic responses have to be subjected to certain pre-treatment before it can be used for purposes of nonlinearity and damage analysis. To approach this problem this work investigates the effect of some key signal and transformation parameters, such as signal length and sampling frequency as well as the SSA window length on the performance of the methodology itself. The selection of these parameters has a direct influence on the damage sensitivity and the accuracy of the methodology. The variation of these parameters can produce radical changes on the clustering effect of the methodology and it is demonstrated that this might affect the results interpretation.

CORE (COnnecting REpositories)