The ice accrued on airfoils can be quite complicated geometrically. Because of this complexity, either unstructured grids or multi-block structured grids are needed. For structured grids, the convergence rate for multi-block grids can be slower than those for single-block grids. This paper examines the convergence rate of different blocking strategies. Blocking strategies evaluated include location of block boundaries in relation to flow features and thickness of the wrap-around block to resolve the viscous region next to solid surfaces. Three ways were used to generate different multi-block grids: (1) partitioning an initially single-block grid, (2) repartitioning an initially multi-block grid, and (3) overlapping different grids. The NLF0414 airfoil with the 623 ice, a moderately complicated ice shape for which single-block grids can still be generated, is used to evaluate the different blocking strategies. For all strategies investigated, two angles of attack were simulated, one low (2.2) and one near stall (5.2). The flow was modeled by the ensemble-averaged compressible Navier-Stokes equations, closed by the shear-stress-transport turbulence model in which the integration is to the wall. All solutions were generated by using the NPARC WIND code. _____________ * Research Associate. ** Graduate Student. # Professor. Associate Fellow AIAA. + Aerospace Engineer, Inlet Branch. Member AIAA. Copyright © 2002 by the American Institute of Aeronautics and Astronautics, Inc. No copyright is asserted in the United States under Title 17, U.S. Code. The U.S. Government has a royaltyfree license to exercise all rights under the copyright claimed herein for government purposes. All other rights are reserved by the copyright owners. INTRODUCTION Ice accrued on airfoils can have catastrophic effects on the aircraft’s aerodynamics. With ice, not only is lift reduced, but also stall occurs at markedly lower angles of attack. Thus, it is important to understand how different ice shapes that can accrue on airfoils affect the aerodynamics of airfoils as a function of angle of attack. Effects of icing on aerodynamics can be studied by flight tests, wind-tunnel tests, and computational fluid dynamics (CFD) simulations. Flight tests are the most realistic but expensive. Tests in wind tunnels offer the advantage of a controlled environment, but there are problems in reproducing actual flight conditions and scaling certain geometric and operating parameters. CFD offers the advantage of low cost and the ability to simulate realistic conditions. However, the accuracy of CFD hinges on the quality of the grid system in resolving the geometry and the flow and the “appropriateness” of the turbulence model in capturing the relevant physics. On grid generation, Chi, et al. presented a number of methods to generate high-quality singleand multi-block structured grids for complicated ice shapes. For moderately complicated ice shapes, single-block grids are possible. But, for really complicated ice shapes, multiblock grids are needed. In their paper, they demonstrated that a thick wrap-around grid is needed to ensure that grid points clustered next to solid surfaces do not propagate into the interior of the flow domain. They also suggested using a transition layer next to solid surfaces that are rough and jagged to confine and smooth the effects of the irregular geometry on the grid. Though Chi, et al. were able to generate high quality multi-block grids for highly complicated ice shapes, convergence rate to steady-state solutions for some multiblock grids were found to be slow, slower than problems for which single-block grids could be generated. The objective of this study is twofold. First, compare the 20th AIAA Applied Aerodynamics Conference 24-26 June 2002, St. Louis, Missouri AIAA 2002-3049 Copyright © 2002 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved. convergence rate of singleand multi-block grids on a consistent basis. Second, examine the effects of blocking strategy on the convergence rate of multi-block grids. Blocking strategies evaluated include location of block boundaries in relation to flow features and thickness of the wrap-around block to resolve the viscous region next to solid surfaces. Three ways were used to generate different multi-block grids: (1) partitioning an initially single-block grid, (2) re-partitioning an initially multiblock grid, and (3) overlapping different grids. Method 1 enables a direct comparison between the convergence rates of singleand multi-block grids since the overall grid is identical. Methods 2 and 3 enable a comparison of different blocking strategies.