The application of Maxwell elements for modeling, identification and analysis of passive and semi-active vibration systems.

摘要

Dynamic analysis and parameter identification of a single mass elastomeric isolation system represented by Maxwell model is examined using both analytical and experimental approaches in this dissertation. Influences that the stiffness and damping values of the Maxwell element have on natural frequency, damping ratio and frequency response are uncovered and three unique categories of Maxwell-type elements are defined. It is revealed through analytical examples that Maxwell models consisting of two Maxwell elements can accurately replicate the dynamic behavior of Maxwell systems having two or more Maxwell elements. Two parameter identification methods are developed for identifying Maxwell models from measured frequency response spectra. To experimentally evaluate the analytic results, three different commercial rubber mounts are considered. For all three rubber isolators, it is shown that Maxwell models with two Maxwell elements can accurately represent the measured static and dynamic characteristics of the real elastomeric isolation systems.; Aeroelastic aircraft wings are the structures which have variable natural frequency and damping ratio as flight parameters change. Serious vibration inhibits the flight at high airspeed conditions. In this study, the dynamic analysis of aeroelastic aircraft wings reveals that a DVA (dynamic vibration absorber) with tunable stiffness and damping parameters can effectively suppress vibration over variable airspeeds in the presence of broadband external disturbance. Since tunable stiffness components are not yet well developed, another configuration of a semi-active DVA having only one tunable damping component is designed. Dynamic analysis reveals that the performance of this semiactive DVA is very close to the DVA having both tunable stiffness and damping components. Two control methods are developed for the semi-active DVA. The first control method is based on the measured airspeed. It works well if the air density is constant during the flight. The second method, a neural-network based controller, is formulated directly in terms of ready measured normalized vibration response spectra. It works well with time-varying airspeed and air density. Both methods are based on measured data and do not require prior knowledge of the plant mathematic model.