Assessment of geometrical variability effects through a viscous through-flow model: sensitivity analysis of geometrical variabilities on performance of axial-flow compressor blades

Authors

Arnaud Budo  - University of Liège, Belgium
Vincent Terrapon - University of Liège, Belgium
Maarten Arnst - University of Liège, Belgium
Koen Hillewaert - University of Liège, Belgium
Jules Bartholet - Safran Aero Boosters, Belgium

Abstract

The aircraft engine market is characterized by a constant search for the compromise between very high performance, and manufacturing cost. The manufacturing tolerance of the blade, and in particular the leading edge, is a good example, as it may have a strong impact on performance and range. It is of utmost importance to be able to predict the effects of tolerances on the performance dispersion, within a time scale compatible with the design process. The computational efficiency of through-flow solvers is especially relevant for parametric studies during design improvement phases in which many parameters or geometrical variabilities can be involved. It is however known that such an averaged two-dimensional CFD modelling is fundamentally limited in its ability to represent turbomachinery flow features, due to either the intrinsic nature of the through-flow equations or the approximations and assumptions made in the closure models (blade forces, ...). Specifically, a through-flow model, even based on Navier-Stokes equations, cannot predict the prerotation of the flow upstream of a blade row due to the axisymmetric assumption on which the model is based. In the case of the analysis mode, the flow angles in the bladed region are imposed by a blade flow deflection force. Therefore a discontinuity occurs at the leading edge if the flow incidence is not aligned with the mean blade camber. This discontinuity leads to a sudden and non-physical increase of entropy and a spurious loss generation. A common fix consists in modifying the blade angle in the leading-edge regions so that it adapts to the incident flow. As a result, the blade geometry is locally modified by this numerical artifact. However, geometrical variability precisely near the blade leading edge has a significant impact on the performance. The goal is thus to assess through a sensitivity analysis how to represent this uncertainty in the through-flow model and how the modelling aspect impacts the variability propagation. In order to study this question, we consider viscous through-flow simulations of two different geometries: the axial compressor stage CME2 and a modern highly loaded multi-stage axial low-pressure compressor. Geometrical variabilities are introduced on the blade geometries to assess the model ability to capture the resulting performance variability with respect to steady 3D RANS computations. Finally, alternative definitions of the incidence correction are discussed in order to minimize the blade geometry smoothing.