Detection of system changes due to damage using a tuned hyperchaotic probe

Abstract

This study explores the use of a hyperchaotic signal as an excitation to probe a system fordynamic changes induced by damage events. In chaotic interrogation a deterministicchaotic input (rather than the more commonly employed stochastic white noise input) isapplied to the structure and the dynamic response is mined for features derived fromits state space reconstruction. The steady-state chaotic excitation is tuned toexcite the structure in a way that optimal sensitivity to dimensionality changes inthe response may be observed, resulting in damage-sensitive features extractedfrom the resulting attractors. The enhanced technique proposed in this paperexplores a hyperchaotic excitation, which is fundamentally new in its use as anexcitation. Hyperchaotic oscillators have at least two Lyapunov exponents, incontrast to simple chaotic oscillators. By using the Kaplan–Yorke conjecture andperforming a parametric investigation, the steady-state hyperchaotic excitation istuned to excite the structure in such a way that the optimal (as will be defined)dimensionality of the steady-state response is achieved. A feature called the ‘averagelocal attractor variance ratio’ (ALAVR), which is based on attractor geometry, isused to compare the geometry of a baseline attractor to a test attractor. Theenhanced technique is applied to analytically and experimentally analyze theresponse of an eight-degree-of-freedom system to the hyperchaotic excitation forthe purpose of damage assessment. A comparison between the results obtainedfrom current hyperchaotic excitation versus a chaotic excitation highlights thehigher damage sensitivity in the system response to the hyperchaotic excitation.