Next: Results - 2-1/2 Stage Up: INVESTIGATION OF AIRFOIL CLOCKING Previous: Algorithm

# Geometry and Grid

The test article is a þstage compressor geometry which models the midspan geometry of an experiment by Dring [10]. The experimental configuration consists of an inlet guide vane (IGV) followed by two rotor/stator pairs. There are 48 IGVs while each of the other rows contains 44 airfoils. It would be prohibitively expensive to compute the flow through the entire 224 airfoil system, so for this computation, the number of IGVs has been reduced to 44. The IGVs have been rescaled by a factor of 48/44 in order to maintain the same blockage as in the experiment. The flow has been computed only through one passage and periodicity has been used to model the other 43 passages.

The axial gaps between airfoil rows in the computational configurations were chosen to be 20%, 35% and 50% of the average of the axial chord of all airfoils in the compressor, and identically matched the axial gaps used in Gundy-Burlet [5]. The 50% axial gap case closely approximates the axial gaps found in the experimental configuration. The circumferential positions of the first-stage rotor (rotor-1) relative to the second-stage rotor (rotor-2) and the first-stage stator (stator-1) relative to the second-stage stator (stator-2) were not documented in the experiment. In the calculations, the rotors are aligned circumferentially. The full computational model with all 8 separate stator displacements in terms of percentage of pitch is shown in Figure 1. The zero displacement position is defined as the one in which the stators are aligned circumferentially. The other positions are evenly spaced at 12.5% of the pitch apart.

A zonal grid system is used to discretize the flowfield within the þ stage compressor. Figure 2. shows the zonal grid system used for the 0% displacement case. In Fig. 2, every other point in the grid has been plotted for clarity. There are two grids associated with each airfoil. An inner, body-centered "O" grid is used to resolve the flow near the airfoil. The thin-layer Navier-Stokes equations are solved on the inner grids. The grid points of the inner grids are clustered near the airfoil to resolve the viscous terms. The Euler equations are solved on the outer sheared Cartesian "H" grids. The rotor and stator grids are allowed to move past each other to simulate the relative motion between rotor and stator airfoils. In addition to the two grids used for each airfoil, there is also an inlet and an exit grid, for a total of 12 grids.

Fine grids are used to obtain detailed data regarding the steady and unsteady flow structure in the compressor. The inner grids are dimensioned (streamwise tangential). The average value of , the non-dimensional distance of the first grid point above the surface, was approximately equal to 1.0 for all five blade rows. The outer grids have an average of 110 points in the axial direction, but they all have 87 points in the circumferential direction. The inlet and outlet grids have 40 and 42 points in the axial direction respectively, for a total of 102064 points in the grid.

Next: Results - 2-1/2 Stage Up: INVESTIGATION OF AIRFOIL CLOCKING Previous: Algorithm

Karen L. Gundy-Burlet
Fri Jul 25 15:51:15 PDT 1997