The flow around a rotorcraft is strongly unsteady and three-dimensional. This complexity is attributable to the unique aerodynamic characteristics of rotary wings and to the reciprocal aerodynamic interaction of the stationary and rotating components that make up a rotorcraft. For a simple helicopter, for example, strong aerodynamic interactions can be observed between the main rotor and the airframe, between the main rotor and the tail rotor, and between the tail rotor and airframe. These interactions dominate the overall flow field and therefore the performance of the helicopter.

Due to the complexity of the flow around rotorcraft, engineers require powerful tools to optimize its performance and stability. Wind tunnel time is difficult to obtain and very expensive, and often of limited effectiveness. Hence, the ability to accurately simulate these complicated flow fields with computational tools becomes increasingly important. Numerical methods can handle many levels of complexity, therefore the computational approach is dictated by the objective of the simulation, which can range from obtaining detailed blade characteristics (stall behavior, noise, FSI, etc.), to cases where the time-averaged cumulative effects of the rotors on each other and the airframe are more than sufficient. For the latter, typical applications would be the prediction of fuselage drag and tail forces for a range of operating conditions, as well the effect of engine exhaust plume distribution and impingement on skin heating and IR-signature prediction. These time-averaged analyses are carried out using reduced-order models such as the Virtual Blade Model.

Historically, in time-averaged simulations, the rotors and their blades are replaced by rotor disk surrogates of equal diameter, or actuator disks, that are coupled with three-dimensional Navier-Stokes or Euler solvers. Since the rotors and their blades are not modelled directly, the computational grids are much smaller, less complex and require significantly less mesh generation effort. With the time-averaged approach, the computational time is also drastically reduced.

Two distinct actuator disk approaches exist. The pressure-disk rotor model simulates helicopter rotors or propellers in a time-averaged manner using a disk composed of two surfaces, one side representing an outflow boundary, the other an inflow boundary. The disk does not need to have a finite thickness, and the grid nodes of the two surfaces need not be paired and coincident. A pressure jump, function of radius and azimuth, is imposed across the disk, subject to the constraint that mass be explicitly conserved through the disk surfaces.

Zori et al.6 (References) developed an alternate technique that replaces the rotor system with momentum sources acting on a one-cell-thick actuator disk that is an integral part of the mesh and does not have inflow and outflow boundaries. The advantage of this approach is that mass is automatically conserved across the disk. Both methods compute the rotor forces according to the Blade Element Theory, using lift and drag coefficient look-up tables for the stacks of airfoils that represent the blades.

In general, for rotorcraft, a rotor operates at the desired thrust and zero moments about the hub. These targets can be met by varying the collective and the cyclic blade pitch angles through user-input or by a trimming algorithm included in the rotor disk model. Typically, the relationship between the thrust coefficient and the collective pitch angle, and between the hub moments and the cyclic pitch angles is assumed to be linear, making the trim routine easy to implement but numerically unstable. The lack of a robust trim routine historically limited the simulation to single rotor configurations, therefore ignoring the tail rotor. Only recently, Yang et al.6, suggested an automatic trim routine based on a Newton-Raphson iterative method to account for the non-linear relation between blade pitch and rotor performance. The current implementation of Fluent VBM is based on the models developed by Zori et al. and Yang et al. (References), with additional enhancements.

Fluent VBM allows for the specification of rotor blades represented by stacks of 2D airfoil sections varying in twist, chord and airfoil type. The airfoil look-up tables, containing lift and drag coefficients versus angle of attack, can also be functions of Mach and Reynolds number, allowing the accurate treatment of both low- and high-speed regimes. Furthermore, the code can handle up to twenty-five individual rotors simultaneously, permitting simulations of complete rotorcraft with both main and tail rotors, or other configurations, such as quadcopters, quad-tiltrotors, or other exotic multi-rotor configurations. Simulations of multiple rotorcraft operating in close proximity are also possible. Furthermore, in the current implementation, the rotor disks can also be meshed with hybrid unstructured grids, simplifying mesh construction for multiple rotors in close proximity and with convenient individual local mesh clustering.