A reduced order model of disk and blade modes interaction

Gas turbine engines are constructed from a series of bladed disks. The vibratory response of these structures is particularly sensitive to small blade to blade frequency variations. The blade frequency variations are caused during manufacturing and may change through use over time. This phenomenon, known as mistuning, can cause failure from high cycle fatigue. The modes of a bladed disk are often categorized by the location of strain energy within the structure. Blade modes have strain energy located mostly in the blades and disk modes have strain energy located mostly in the disk. Families of blade modes are modes with similar frequencies and that have similar deformation patterns. When blade modes and disk modes are close to one another in frequency, they can undergo modeshape changes through a phenomenon known as veering.;Numeric and experimental benchmark models of mistuned bladed disk systems are analyzed in this thesis. A mass-spring representation of a bladed disk was used to demonstrate the identification and prediction capabilities in a simple system. A finite element model of an actual compressor was used to show those capabilities for a more complex, realistic geometry. In addition to these purely numeric models, a bladed disk was constructed and experimentally measured in multiple mistuned states. The measured modes and natural frequencies from these experiments were used as an experimental benchmark to corroborate the validity of the approach developed in this thesis.;A new approach for modeling the vibration of mistuned bladed disks undergoing a veering between disk and blade modes is presented in this thesis. This approach utilizes a reduced order model that includes the effects of veering between a single disk mode pair and a family of blade modes. This model has the ability to use experimental data to identify the mistuning in the blades and the disk, as well as a new type of mistuning that is associated with the cross-product of the blade and disk modes. Using these mistuning parameters as input, the model can also be used to predict the mistuned modes and natural frequencies of bladed disks.