Nonlinear Reduced Order Models and Damage Detection in Systems with Complex Geometry.

[摘要] Recently, the interest in monitoring the structural health of dynamic systems has beenincreasing. Especially large and complex air and space structures need monitoring systemsto help ensure the safety and reliability of the system during operation. To monitorthe structural health of the system, vibration characteristics such as amplitude and frequencyare often used. For example, vibrations measured in turbine engine rotors (usingtip timing) can be used to examine the health of the system. Turbine engine rotors experiencecomplex dynamics induced by aeroelastic coupling, multi-stage structural coupling,vibration localization, and nonlinearity (for example caused by damage). Thus, to accuratelyinterpret the measurement data, efficient models to capture the effects of thosefactors are needed.Finite element (FE) models are often used to predict vibrations of structural systems.For low dimensional systems, full FE models can be used. For high dimensional systems,the computational cost of analyzing full FE models can often be prohibitive. To circumventthis difficulty, many methods for creating efficient reduced order models (ROMs)have been presented for various systems. However, most of these methods are focused ondeveloping ROMs for linear systems by using linear transformations such as componentmode synthesis.This work presents highly efficient novel methods for predicting nonlinear vibrationsand for detecting damage in dynamic systems with complex geometry. Efficient ROMsfor analyzing the effect of aeroelastic coupling and cracks on the vibration of turbine enginerotors are introduced followed by a novel method to detect cracks in bladed disks.Also, a new and efficient method to approximate the vibration amplitude at the resonantfrequencies of dynamic systems with piecewise-linear nonlinearity is presented. In addition,efficient ROMs designed to capture the effects of (nonlinear) Coulomb friction atinterfaces is presented and applied to combustors used for power generation.

[发布机构] University of Michigan

[效力级别] Mechanical Engineering

[关键词] NONLINEAR REDUCED ORDER MODELS AND DAMAGE DETECTION IN SYSTEMS WITH COMPLEX GEOMETRY;Mechanical Engineering;Engineering;Mechanical Engineering