Detailseite
Projekt Druckansicht

Mistuning mit Aero-Kopplung II (Mistuning aerodynamisch und strukturmechanisch gekoppelter Beschaufelungen)

Fachliche Zuordnung Mechanik
Strömungs- und Kolbenmaschinen
Förderung Förderung von 2013 bis 2021
Projektkennung Deutsche Forschungsgemeinschaft (DFG) - Projektnummer 239841452
 
Erstellungsjahr 2020

Zusammenfassung der Projektergebnisse

The simulation tool ROMI has been supplemented with additional interblade coupling mechanisms for the mistuned vibration prediction. In the context of forced response simulations in a limited frequency range, the effect of frequency variations on the aerodynamic coupling coefficients has been addressed. The results indicate a frequency dependence of the intra-modal influence coefficients between directly adjacent blades. The corresponding trend for the off-diagonal inter-modal coefficients is more complicated and requires further investigation. Furthermore, an interpolation of the frequency-dependent coefficients during a forced response simulation is appropriate in a narrow frequency range around the analyzed resonance. On the structural side, the modeled friction joints have been complemented with nonlinear contacts between the blade roots and the disk. The implementation allows the consideration of simple dovetail joints as well as fir-tree roots with multiple flanks. Preliminary studies with either mistuned root contacts or mistuned blade frequencies illustrate the significance of the nonlinear contact forces. It has been shown that the effect of irregular contact normal forces across the blades on the overall response amplification strongly depends on the nominal preload level at the root. Similarly, the response localization due to a frequency mistuning of the individual blades depends on the disk-blade coupling as well as on the friction damping which are both influenced by the root preload. To increase the efficiency of the nonlinear forced response simulations, different concepts for the reduction of the interface degrees of freedom have been analyzed. The accuracy of a rigid interface approximation has been compared to a higher-order approach based on Legendre polynomials. Since the former represents a special case of the latter, the polynomial approach outperforms the rigid interface approximation in terms of prediction accuracy. Both reduction methods have been applied to a large scale model of an axial compressor disk with dovetail joints. Following the supplementation of the reduced order model by aerodynamic coupling effects and nonlinear contact interfaces, two measurement campaigns have been conducted to allow a comparison between numerical reduced order model (ROM) results and measurement data. Each measurement focused on either aerodynamic coupling or friction contacts in order to address their effects separately from each other. Regarding the aerodynamic coupling effects in the presence of mistuning, forced response measurements of two axial compressor blisks have been carried out at the high-speed axial compressor rig of the Institute of Turbomachinery and Fluid Dynamics (TFD). While the reference blisk consists of 24 nominal blade geometries, the modified blisk features three blades with blend repairs. Both, blade tip timing measurements and ROM simulations for various engine orders, indicated a similar trend for the maximum response amplitude. The measurements indicated a response reduction while the ROM results revealed a slightly increased amplitude compared to the nominal reference. The reason for this may be related to small unknown mistuning of the reference blisk. Concerning the nonlinear contact interfaces, a bladed disk with 16 blades with dovetail joints has been designed and put into operation. The bladed disk is operated in the rotational test rig at the Institute of Dynamics and Vibration Research (IDS) under vacuum conditions. The individual mistuning of each blade after manufacturing has been identified experimentally. Furthermore, additional mistuning masses have been applied to the blade tips in order to realize an intentional A/B mistuning pattern. A comparison of the numerical ROM simulations and the measurement data showed a good agreement in terms of response amplitudes for various engine orders and rotational speeds. This holds true for both the stochastically and intentionally mistuned configuration. Regarding the resonance frequencies, the ROM predicted slightly smaller frequency values due to the limited amount of contact nodes used for the simulations.

Projektbezogene Publikationen (Auswahl)

 
 

Zusatzinformationen

Textvergrößerung und Kontrastanpassung