CFM 2019

3D effects on flutter prediction in aeronautic compressor
Pierre Duquesne  1, *@  , Bruno Mahieux  2@  , Stéphane Aubert  1@  , Pascal Ferrand  1@  
1 : Laboratoire de Mecanique des Fluides et d'Acoustique  (LMFA)
CNRS : UMR5509, Université Claude Bernard - Lyon I (UCBL), Ecole Centrale de Lyon, Institut National des Sciences Appliquées [INSA] - Lyon
2 : Safran Aircraft Engines  (SAE)
Aucune
* : Auteur correspondant

The flutter corresponds to an aerodynamic loading of the structure which amplifies the natural vibration of the structure. In turbomachinery, the flutter appears when the rotation speed match with the frequency of one of the natural blade vibrations. This fluid-structure interaction can lead to the failure of compressor blades. In the design stage, the detection of flutter is estimated in a 2D radial plan (a blade to blade channel, near the shroud) with numerical simulations or with empirical design criteria. The studies on modern designs point to some disparity between these predictions and experimental test. 3D and 2D simulations have been conducted on the rotor (isolated rotor) of a modern design of high pressure axial compressor to investigate the limitations of the 2D approach.

To solve the steady flow, the Reynolds-Averaged Navier-Stokes (RANS) simulations with k-ω turbulence model have been performed with elsA. Next, the blade motion corresponding to the first torsion mode at two nodal diameters has been imposed on the blade. The mode shape, the nodal diameter and the frequency have been extracted from a solid mechanical analysis software. The flow fluctuations have been determined by the time-linearized RANS solver Turb'lin. Two operating conditions have been selected, one at nominal speed and the second at part speed regime. At nominal speed, a strong shockwave chokes the blade to blade channel, from the blade half-height to the shroud. At part speed, the shockwave only happens on the suction side.

Analysis has been based on the comparison of the work exchanged between the 2D and the 3D calculations. Then, the work has been discretizing between the contribution generates by the steady pressure and the blade motion and the contribution induces by the pressure fluctuation and the blade motion.

The 3D calculation confirms that the height of the 2D-extraction (at 80% heights in this case) is representative of the work exchange of the entire blade. Despite this, for the operating point at nominal speed, the amount of work induced by the pressure fluctuation is different between the 2D or 3D calculation. This unsteady work representing a major part of the total work, so pressures wave propagation has been investigated using pressure fluctuation mapping. At nominal speed, pressure waves present propagation in radial direction. Contrary, at part speed pressure waves only propagate in axial direction. The radial propagation cannot be captured with a 2D calculation and explaining the disparity of results. This effect seems to be more dominant than the projections effect needed to perform a 2D calculation.

In the blade to blade channel, regressive pressure waves from the trailing edge travel up to the shock wave. Approaching the shockwave, pressure wave amplitudes increase and induce important pressure fluctuations. In the 3D calculation pressure waves can escape in radial direction. This mechanism not necessarily induces a stronger pressure fluctuation at the shockwave but induce a temporal shift. The synchronisation of the local pressure force and the motion of the blade being an essential flutter parameter. A different timing can induce a different amount, or worst reverse the direction, of the work exchange.


Personnes connectées : 59