The CSIRís Defence Aeronautics CFD group is directing substantial effort in simulating the flow field around helicopters in hover and forward flight. At this stage CFD models of two well-known helicopters, viz. Rooivalk Combat Helicopter and the Oryx Transport and Medium Lift Helicopter, have been developed and simulation results have already been used in the design and countermeasure environments. The CD adapco Groupís CFD code, STAR-CD, and in particular itís automatic meshing tool, pro-am, have been invaluable tools in the development of this technology.
If one considers the nature of the flow field around a helicopter in hover or trimmed forward flight and the complex geometry of the entire aircraft, one can understand the extreme demands placed on the CFD software and hardware. While it is possible to simulate a hovering helicopter using steady state simulation techniques, a transient simulation is required for the helicopter in trimmed forward flight as each rotor blade pitches as it rotates around the rotor shaft through the helicopter azimuth. Thus the motion of each blade has to be specified and controlled individually as the simulation proceeds. Moreover, the tail rotor movement must be specified independently from that of the main rotor. STAR-CDís advanced mesh movement capability enables one to model such a case.
The fuselages are complex, and multi-block mesh generation techniques proved time-consuming and inefficient. Automatic meshing comes to the rescue! Solid models were firstly generated in SolidWorks and surfaces were wrapped around the solid fuselages using a STAR-CD plug-in. These surfaces were conditioned and modified to an acceptable level using pro-amís surface tools. Thereafter, the original surface was expanded by an offset distance to create a subsurface. Using alternating cell classification and local refinement in pro-am, an efficient trimmed-cell mesh, external to the subsurface, was created. The sub-layer, a layer of cells normal to the helicopter surface, fills the space between the original surface and the subsurface, allowing more control over the cell quality adjacent to the helicopter surface and improving turbulence modeling and viscous drag prediction.
Rotor collective, lateral and longitudinal rotor tilt, lateral and longitudinal cyclic settings, and rotor geometry are among the various parameters required to model the main and tail rotors. These parameters, together with the fuselage attitude, are extracted from dynamic flight simulations using CAMRAD J/A and entered into a FORTRAN code that automatically generates the independent rotor grids using block structures, and performs the integration with the fuselage grid. A mesh movement script, with the lateral and longitudinal cyclic pilot inputs, controls the pitch attitude of each main rotor blade while being rotated around the rotor centre through the helicopter azimuth. The same script rotates the tail rotor.
The modeling of the fuselage and the main rotor is being validated with results from tests conducted in the Low Speed Wind Tunnel and 7 metre wind tunnel at the Defence Aeronautics Programme. This intense validation exercise is planned to take place over the next two years with funding from the SAAF and Armscor. However, this CFD Groupsí efforts to date have provided invaluable information and insight for the South African Air Force.
Glen Snedden, head of CFD at CSIR, says, "pro-am's tools for meshing complex aerodynamic bodies coupled with the moving mesh capability of STAR-CD have made it possible for us to achieve excellent results in modeling the flowfield complexities of a helicopter with moving main and tail rotors, even when our hardware resources were more limited."
Velocity magnitude contours for Oryx in hover.
Pressure contours on helicopter surface in forward flight.