Reduced-Order Models For Robust Identification and Damage Modeling of Blisks.

Reduced Order Models;Structural Dynamics;Turbomachinery;Cyclic Symmetry;Vibrations;Mechanical Engineering;Engineering;Mechanical Engineering

Madden, Andrew C.Vlahopoulos, Nickolas ;

University of Michigan

关键词: Reduced Order Models;    Structural Dynamics;    Turbomachinery;    Cyclic Symmetry;    Vibrations;    Mechanical Engineering;    Engineering;    Mechanical Engineering;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/84432/maddena_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

This work introduces novel reduced order modeling techniques for identification ofsmall mistuning and the simultaneous modeling of large and small mistuning in largestructures that are nominally cyclically symmetric. Mistuning, or deviations in cyclicallysymmetric structures, can cause dramatic changes in the forced response of a system.Turbomachinery blisks are an example of cyclic structures that are subject to mistuning.Mistuning can be caused by small or large deviations in manufacturing, wear, or damage.Many accurate reduced order models (ROMS) exist to predict the system response inthe presence of small mistuning. These will be denoted as direct ROMs, or those ROMsthat generate system response predictions given knowledge of the system. However, thegrowing use of blisks, makes it necessary to determine the mistuning levels using systemlevel experimental data of the response. Enabling this is the use of a novel concept developedin this work and denoted as inverse reduced order models (IROMs), which are modelscarefully selected to enhance sensitivity to the parameters that need to be predicted.System modes are categorized and a process of surrogate data is used to systematicallyevaluate IROMs.Many ROMs which model small mistuning in blisks are only accurate for small frequencyranges and are untested in higher frequency ranges. However, it is also necessaryto have a tool to identify mistuning in higher frequency ranges. This is a nontrivial stepbecause blade mode families are no longer isolated and blade topology becomes increasinglycomplicated. An enhanced measurement point selection algorithm integrated with asurrogate data process enables the identification of mistuning at higher frequencies.Large mistuning modeling of blisks using ROMs is an important step to predicting theresponse of the system. The approach in this work makes use of modes solved using cyclicsymmetry analysis to predict the pristine, rogue, and interface portions of the system.This allows for the ROM to be generated using only simple sector level calculations, thusenabling rapid generation of ROMs. This approach can be combined with small mistuningmodels to accurately model both small and large mistuning in blisks.