Forced Response Sensitivity of a Mistuned Rotor From an Embedded Compressor Stage

Abstract

The frequency mistuning that occurs due to manufacturing variations and wear and tear of the blades can have a significant effect on the flutter and forced response behavior of a blade row. Similarly, asymmetries in the aerodynamic or excitation forces can tremendously affect the blade responses. When conducting CFD simulations, all blades are assumed to be tuned (i.e. to have the same natural frequency) and the aerodynamic forces are assumed to be the same on each blade except for a shift in interblade phase angle. The blades are thus predicted to vibrate at the same amplitude. However, when the system is mistuned or when asymmetries are present, some blades can vibrate with a much higher amplitude than the tuned, symmetric system. In this research, we first conduct a deterministic forced response analysis of a mistuned rotor and compare the results to experimental data from a compressor rig. It is shown that tuned CFD results cannot be compared directly with experimental data because of the impact of frequency mistuning on forced response predictions. Moreover, the individual impact of frequency, aerodynamic, and forcing function perturbations on the predictions is assessed, leading to the conclusion that a mistuned system has to be studied probabilistically. Finally, all perturbations are combined and Monte-Carlo simulations are conducted to obtain the range of blade response amplitudes that a designer could expect.

Dynamic Modeling and Projection-Based Reduction Methods for Bladed Disks With Nonlinear Frictional and Intermittent Contact Interfaces

“…Mistuning occurs when there is a variability across the sectors of the blisk in a parameter, which affects its dynamics. Variations in different material properties such as mass, stiffness, and damping as well as in geometry across sectors have been studied in literature [4,5,[8][9][10][11][12][13][14]27,[105][106][107]. The consequence is that the sector-level system matrices are nonidentical across sectors.…”

“…If any sector of the blisk is considered to be geometrically and materially identical to any another, the DoFs of the blisk dynamic may be reduced by a factor of the number of blades by the application of suitable constraints. However, it is well known that even small differences, which exist between different sectors in practice due to manufacturing tolerances, collectively called mistuning, can have significant effects on blisk response amplitudes [4][5][6][7][8][9][10][11][12][13][14]. Mistuning may be small when the variability in parameters between the sectors does not significantly affect the modes of the structure, or large when the response of the mistuned blisk in a certain frequency range cannot be captured by a set of nominally cyclic symmetric or tuned modes whose natural frequencies lie in the corresponding range.…”

Suppression of Nonsynchronous Vibration Through Intentional Aerodynamic and Structural Mistuning

“…This can be either structural, where the natural frequencies of individual blades around the circumference vary, or aerodynamic, with blade-to-blade variations in the (timeaveraged) flow field. This has been extensively studied in the context of flutter and forced response, see for example References [16][17][18] The aim of this paper is to demonstrate and explain the effectiveness of different types of mistuning (structural and aero-dynamic) to prevent convective NSV. This will be achieved by incorporating mistuning in the model developed in Reference [1] and [19], which was previously validated against experimental measurements for the Darmstadt transonic compressor.…”

Analysis of mistuned forced response in an axial high-pressure compressor rig with focus on Tyler–Sofrin modes

“…Such a ROM is traditionally used to study the effect of mistuning on the forced vibration behaviour due to a single engine order excitation acting on a single travelling wave mode (TWM) ( 2 , 7 , 13 , 14 ) . The derived frequency response functions (FRFs) and blade individual vibration amplitudes can then be compared with those obtained in measurements – sometimes with unsatisfying agreement on some of the blades and unexpected nodal diameter content in the response ( 13 ) . This might be due to the simplification of the force vector by restricting it to act on a single nodal diameter which does not hold in general – especially in a multistage engine.…”