Computational Modeling of Rotor Blade Performance Degradation Due to Ice Accretion

Open Access
Author:
Brown, Christine Meredyth
Graduate Program:
Aerospace Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
August 29, 2013
Committee Members:
  • Kenneth Steven Brentner, Thesis Advisor
Keywords:
  • icing
  • CFD
  • turbulence modeling
  • mesh generation
Abstract:
Performance degradation of rotor blade ice shapes was computationally investigated for several icing conditions. The ice shapes were accreted in the Adverse Environment Rotor Test Stand (AERTS) at Pennsylvania State University, and digitized to provide 2D slices of the iced blade. Hybrid, viscous sub-layer resolved grids were created in order to capture these ice features. Several sub-layer turbulence models were applied. For 2D cases, drag predictions obtained with the NPHASE-PSU Computational Fluid Dynamics (CFD) code compare favorably with drag data from several data sources including the AERTS experimental facility. Iced airfoil drag is primarily due to form drag associated with the bluff effective leading edge ice shapes, effects which are well captured. However, predicted lift values deviate signi cantly from the experimental data at high angles-of-attack, consistently underpredicting stall onset and high-alpha-lift. Parametric exploration of the causes of these discrepancies was pursued. Specifcally, experiments performed in NASA Glenn's Icing Research Tunnel on two additional airfoil geometries with accreted glaze ice shapes were modeled. These studies show that 2D RANS turbulence modeling consistently leads to this same inaccurate lift prediction behavior, and that this observation is consistent with other studies that have appeared in the literature. These shortcomings are hypothesized to arise due to the well known challenges that 2D RANS turbulence models have in adverse pressure gradient flows, stagnating flows, and flows with strong curvature and also due to the inherent three-dimensionality of the ice shape and the unsteady recirculation regions aft of the sharp ice shape separation points. In order to explore the role of 3D and unsteady flow effects, a CT scan of one of the full 3D AERTS ice molds was used to create a large 3D grid. NPHASE-PSU was run using an Implicit Large Eddy Simulation (ILES) turbulence model. These results indicate that incorporation of three-dimensionality and large scale unsteadiness (upper inertial range) are important in capturing the aerodynamics in these systems.