Probabilistic CFD Analysis of a High-Pressure Compressor Under Consideration of Manufacturing and In-Service Variability

Abstract

Abstract The impact of geometric variability of compressor airfoils due to manufacturing scatter and operational effects on the aerodynamic performance is a well-addressed topic in the research field of turbomachinery. Thereby, the airfoils’ leading edge (LE) shapes are highly connected to their aerodynamic behavior, affecting the boundary layer development and potential shock systems. The impact of such real geometry effects can be evaluated utilizing probabilistic methods. In previous probabilistic computational fluid dynamics (CFD) analyses, the LEs are usually modeled as semi-circles or semi-ellipses, whereas their noticeable deviation from such simplified shapes is given comparatively little attention. To overcome this shortfall a Monte Carlo simulation based probabilistic CFD analysis of a 10.5 stage high-pressure compressor is conducted under consideration of variable rotor geometries. These are described by a parametric model accounting for the airfoils’ general- and the LE shapes’ specific geometric variability. A database with more than 1000 optically measured blades is utilized for deriving median models (MMs) and the statistical distributions of the deviation from these MMs. Additionally, the assessment of the parameterization quality is based on these measurements. The CFD results are post-processed using the coefficient of importance, which allows the identification of important probabilistic input parameters and prominent rotors. The probabilistic results highlight the significance of the LE shapes’ variability. Furthermore, the impact of the sample size on the probabilistic results as well as the importance of characteristic differences within the underlying MMs are discussed.