0
Research Papers

Mistuned Forced Response Predictions of an Embedded Rotor in a Multistage Compressor

[+] Author and Article Information
Fanny M. Besem

Department of Mechanical Engineering,
Duke University,
Durham, NC 27708
e-mail: fanny.besem@duke.edu

Robert E. Kielb

Department of Mechanical Engineering,
Duke University,
Durham, NC 27708
e-mail: rkielb@duke.edu

Paul Galpin

ANSYS Canada Ltd.,
Waterloo, ON N2J 4G8, Canada
e-mail: Paul.Galpin@ansys.com

Laith Zori

ANSYS Inc.,
Lebanon, NH 03766
e-mail: Laith.Zori@ansys.com

Nicole L. Key

Zucrow Laboratories,
School of Mechanical Engineering,
Purdue University,
West Lafayette, IN 47907
e-mail: nkey@ecn.purdue.edu

1Corresponding author.

2Present address: Numeca, San Francisco, CA 94109.

Contributed by the International Gas Turbine Institute (IGTI) of ASME for publication in the JOURNAL OF TURBOMACHINERY. Manuscript received October 17, 2015; final manuscript received November 4, 2015; published online February 9, 2016. Editor: Kenneth C. Hall. Review conducted by David Wisler.

J. Turbomach 138(6), 061003 (Feb 09, 2016) (10 pages) Paper No: TURBO-15-1228; doi: 10.1115/1.4032164 History: Received October 17, 2015; Revised November 04, 2015

This paper covers a comprehensive forced response analysis conducted on a multistage compressor and compared with the largest forced response experimental data set ever obtained in the field. The steady-state aerodynamic performance and stator wake predictions compare well with the experimental data, although losses are underestimated. Coupled and uncoupled unsteady simulations are conducted on the stator–rotor configuration. It is shown that the use of a decoupled method for forced response cannot yield accurate results for cases with strong inter-row interactions. The individual and combined contributions of the upstream and downstream stators are also assessed. The downstream stator is found to have a tremendous impact on the forced response predictions due to the constructive interactions of the two stator rows. Finally, predicted mistuned blade amplitudes are compared to mistuned experimental data. The average amplitudes match the experiments very well, while the maximum response amplitude is underestimated.

Copyright © 2016 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

Drawing of the compressor located at the Zucrow Laboratories in the Purdue University. The air flow comes from left to right. The number of blades is indicated above each row. Adapted from Ref. [13].

Grahic Jump Location
Fig. 2

Campbell diagram for the second-stage rotor. The crossing of interest in this paper occurs at 3700 rpm, at the crossing between the 44/rev coming from the upstream and downstream vanes (inlet guide vanes, first-, and second-stage stators) and the first torsion frequency.

Grahic Jump Location
Fig. 3

Experimental and numerical speedline at 74% corrected speed. The experimental speedline at 68% corrected speed is also shown for completeness.

Grahic Jump Location
Fig. 4

Contours of relative Mach number throughout the compressor at 20% span for the nominal operating point at 74% corrected speed

Grahic Jump Location
Fig. 5

Comparison of the stator 1 wake profile at 50% and 80% span between CFD and experiments at 74% Nc

Grahic Jump Location
Fig. 6

Fourier transform of the stator 1 wake at 50% and 80% span for the CFD and the experiments at 74% corrected speed. The wider bars represent the harmonic amplitudes from the CFD wake at 50% span, while the narrower bars are from the experimental data at 50% span.

Grahic Jump Location
Fig. 7

Surface mesh on the vanes, blades, and hub of the model used in the TT transient simulations. Three stator 1 vanes and two rotor 2 blades are modeled.

Grahic Jump Location
Fig. 8

Modal force amplitude density on the pressure side (PS) and SS of the rotor blade for the transient TT simulation with three stator 1 vanes and two rotor 2 blades

Grahic Jump Location
Fig. 9

Modal force amplitude density on the rotor 2 surface for the decoupled two-row configuration. PS and SS signify pressure and suction side, respectively.

Grahic Jump Location
Fig. 10

Total pressure contours in the stationary frame of reference at the inlet of rotor 2. (a) Boundary profile extracted from the whole compressor steady CFD and imposed at the rotor passage inlet. (b) Boundary profile calculated at the interface between stator 1 and rotor 2 in the two-row transient TT configuration with multiple passages per row.

Grahic Jump Location
Fig. 11

Amplitude of the surface pressure first harmonic at 50% span of the rotor blade. The one- and two-row configurations with transient TT models are shown, both simulations include two rotor passages.

Grahic Jump Location
Fig. 12

Amplitude of the surface pressure first harmonic at 90% span of the rotor blade. The one- and two-row configurations with transient TT models are shown, both simulations include two rotor passages.

Grahic Jump Location
Fig. 19

Comparison of the envelope of the un-normalized average blade response amplitude over the frequency sweep through the resonant frequency between experiments and predictions

Grahic Jump Location
Fig. 18

Variations from the mean of the experimentally measured natural frequencies for the 33 rotor blades. The mean natural frequency is 2723 Hz.

Grahic Jump Location
Fig. 17

Pressure first harmonic amplitude at 90% span of the rotor blade for the two- and three-row coupled transient configurations

Grahic Jump Location
Fig. 16

Pressure first harmonic amplitude at 50% span of the rotor blade for the two- and three-row coupled transient configurations

Grahic Jump Location
Fig. 15

Contours of modal force amplitude density on the suction and PS of the rotor 2 blade for the three-row configuration. PS and SS signify pressure and suction side, respectively.

Grahic Jump Location
Fig. 14

Contours of total pressure in the stationary frame of reference at the interface between stator 1 and rotor 2. Three vanes and two blades are modeled in the transient TT simulation. The reader is forward looking aft.

Grahic Jump Location
Fig. 13

Surface mesh on the vanes, blades, and hub of the model used in the three-row transient simulations. Three stator vanes are modeled on stator 1 and 2 and two rotor 2 blades are modeled.

Grahic Jump Location
Fig. 20

Comparison of the envelope of the predicted un-normalized maximum blade response amplitude over the frequency sweep through the resonant frequency for all three configurations

Grahic Jump Location
Fig. 21

Comparison of the envelope of the un-normalized maximum blade response amplitude over the frequency sweep through the resonant frequency between experiments and predictions. The predictions use the three-row coupled transient analysis and the mistuning predictions from Ref. [25]. The plain black curve represents the tuned predictions using the three-row configuration.

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In