Study of Oscillating Blades from Stable to Stalled Conditions 1 CFD Lab, Department of Aerospace Engineering, University of Glasgow 2 Volvo Aero Corporation 3 Rolls-Royce Plc S. Svensdotter 3, G. Barakos 1 and U. Johansson 2

Motivation Turbomachinery blades flutter Turbomachinery blades flutter During flutter blades may break During flutter blades may break Implications on the safe operation of the engine Implications on the safe operation of the engine

Flutter Structural vibration involving bending and twisting Structural vibration involving bending and twisting A result of interactions between aerodynamics, stiffness and inertial forces A result of interactions between aerodynamics, stiffness and inertial forces Can be experienced on all flexible structures Can be experienced on all flexible structures

Engine Working Line

Objectives Study a simple peculiar case of flutter Study a simple peculiar case of flutter Use CFD to assess the quality of experimental data Use CFD to assess the quality of experimental data Shed some light in the argument raised about this particular flutter case Shed some light in the argument raised about this particular flutter case

Background Three key publications on the subject Three key publications on the subject  (1) Parametric Study of the Pressure Stability of an Oscillating Airfoil from Stable to Stalled Flow Conditions (S. Svensdotter, U. Johansson, T. Fransson)  (2) Boundary-Layer Transition, Separation and Reattachment on an Oscillating Airfoil (T. Lee, G. Petrakis)  (3) An Experiment on Unsteady Flow Over an Oscillating Airfoil (L. He, J.D. Denton)

Experimental Method Equipment Equipment  Symmetrical 2D NACA 63A006, chord length 80mm, span 150mm  13 pressure transducers on the suction surface of the blade  Pitching axis at 43% chord  Performed in a wind tunnel with test section 150mm by 180mm

Experimental Method Test program and data interpretation Test program and data interpretation  High oscillating frequencies (up to 210Hz)  Inlet Mach number was 0.5  Reynolds number was  Unsteady pressure signals were analysed in terms of amplitude of perturbation and the phase difference between the pressure signal and blade motion

VAC wind tunnel air system AWAITING PICTURE

Test rig

Test equipment

Amplitude and Phase, 0° Incidence

Amplitude and Phase, 6° Incidence

Summary of measurements Highly non-linear behaviour at and above static stall angle Highly non-linear behaviour at and above static stall angle Below stall angle Below stall angle  Airfoil damped  Increase amplitude => increased excitation  Increase to 210Hz => blade excited Above stall angle Above stall angle  Airfoil excited  Increased amplitude => decreased excitation  Increase to 210Hz => blade damped 210Hz phase shift possibly due to lagging LE separation vortex or migration of stagnation point 210Hz phase shift possibly due to lagging LE separation vortex or migration of stagnation point

Summary Of Findings (2) Transition & Separation, Relaminarisation & Reattachment delayed with increasing reduced frequency (T. Lee, G. Petrakis). (2) Transition & Separation, Relaminarisation & Reattachment delayed with increasing reduced frequency (T. Lee, G. Petrakis). (3) Increasing frequency delays dynamic stall (L. He, J.D. Denton) (3) Increasing frequency delays dynamic stall (L. He, J.D. Denton)

Analysis Use CFD to simulate the experiment Use CFD to simulate the experiment Used the University of Glasgow PMB code to analyse the data Used the University of Glasgow PMB code to analyse the data Cross-plotted the CFD results and the experimental results in order to make a comparison Cross-plotted the CFD results and the experimental results in order to make a comparison

CFD Grid 1 st block 41*85*2 3 rd block 41*85*2 2 nd block 222*85*2

CFD results on the pressure field

Table of Cases Frequency60Hz110Hz210Hz Inlet mach No Inlet stagnation temperature 280K280K280K Reynolds number

CFD Cases Case A Reference Case 210Hz Case B Free-stream turbulence 1% Case C As Reference at twice the frequency Case D As Reference with free-stream turbulence 0.001% Case E Twice the frequency and turbulence 0.001% Case F Four times the frequency

Time Domain Comparison 60Hz

Frequency Domain Comparison 60Hz

Time Domain Comparison 110Hz

Frequency Domain Crossplots 110Hz

Time Domain Crossplots 210Hz

Frequency Domain Crossplots 210Hz Reference case

Frequency Domain Crossplots 210Hz Case F

Boundary layer behaviour Turbulent Reynolds Number indicating laminar flow up to 30% of the chord Pressure taps indicating transition at ~50% chord

BL trip at 50% chord Mean incidence 0 degrees Amplitude 1 degree Frequency 120 Hz Trip at x/c=0.5 Mean incidence 7 degrees Amplitude 1 degree Frequency 120 Hz Trip x/c=0.5

Mean incidence 7 degrees Amplitude 1 degree Frequency 210 Hz Mean incidence 7 degrees Amplitude 1 degree Frequency 210 Hz Trip x/c=0.5 BL trip at 50% chord

Mean incidence 7 degrees Amplitude 1 degree Frequency 210 Hz Trip x/c=0.5 Mean incidence 4 degrees Amplitude 1 degree Frequency 210 Hz Trip x/c=0.5 BL trip at 50% chord

Mean incidence 0 degrees Amplitude 1 degree Frequency 120 Hz Transition at x/c=0.5 Mean incidence 0 degrees Amplitude 1 degree Frequency 210 Hz Transition at x/c=0.5 BL trip at 50% chord

Conclusions  The 60Hz and 110Hz cases were reasonably the same  210Hz Experiment suggests a change in phase, CFD maintains the same trend, for reference case  BL trip at 50% chord gives phase shift, even for zero incidence  The higher mean incidence the easier to get phase shift  State of boundary layer seems to affect phase  Investigate difference of BL between 110Hz and 210Hz frequences  Further work needs to be carried out

My rough conclusions At high mean incidence phase differences exist at 210 Hz and the trip makes the phase difference more pronounced (ie more probes are at 180 degrees of difference) At high mean incidence phase differences exist at 210 Hz and the trip makes the phase difference more pronounced (ie more probes are at 180 degrees of difference) Reducing the mean incidence and keeping the trip maintains the phase difference at 210 Hz (probes up to the middle of the aerofoil are at 180 degrees of difference) Reducing the mean incidence and keeping the trip maintains the phase difference at 210 Hz (probes up to the middle of the aerofoil are at 180 degrees of difference) The trip at 210 Hz and zero mean incidence results at phase difference of 180 degrees, however, the trip induces some phase difference even at 120 Hz The trip at 210 Hz and zero mean incidence results at phase difference of 180 degrees, however, the trip induces some phase difference even at 120 Hz This may be due to the fixed transition employed and we need a more intelligent way of modelling transition…something that gives different transition locations at different frequencies. This may be due to the fixed transition employed and we need a more intelligent way of modelling transition…something that gives different transition locations at different frequencies. Effect of frequency on the BL stability needs to be examined. Effect of frequency on the BL stability needs to be examined.