Modelling and Simulation of helicopters and tilt-rotors in Vortex Ring State

Progress with computational fluid dynamics based on the Navier-Stokes equations allows for the computation of flows around propellers within realistic time-scales. The most established method for such simulations is to employ the Unsteady Reynolds-Averaged Navier-Stokes method that is a good compromise between accuracy and efficiency. This is because URANS doesn't need to resolve all flow scales which leads to CFD meshes of the size of millions of cells and time steps that are of 1/100 of the chord travel-time of particles. The results can be used for the computation of the mean pressure and velocity fields as well as the slowest frequencies present in the flow. When translated to acoustics, the method can only give tone-noise while the broadband spectrum remains under-resolved.

In terms of geometric complexity, the use of structured or unstructured grids is possible, and by employing overset or sliding meshes, the predictions can account for the interaction between the wing and nacelle of an installed propeller system.

The Helicopter Multi-Block method of Glasgow has all the ingredients for computing the flow around installed propellers based on the URANS approach and it is therefore a good starting point for research work in this area.

On the other hand, the need to resolve more and more harmonics as well as the broad-band part of the acoustic spectrum present in propeller flows, requires more sophisticated techniques that are based on simulation rather than modelling of turbulence. Glasgow has substantial experience with Detached Eddy Simulation for the computation of flows inside weapon bays and helicopter rotors. DES could, in principle, be used for propellers and should allow for the resolution of a large part of the flow spectrum at the expense of more computational resources. For a flow around a propeller blade, grids of the order of 10 million cells should be used with this method and time steps of the order of 1/10000 of the chord travel-time. This of course leads to an increase of the required CPU time by a factor of 100 in comparison to URANS (due to the use of efficient time-integration schemes in the HMB solver). To avoid the penalty associated with this method. The recently-developed Structure-Adaptive-Simulation or SAS should be tested for propeller flows. This method should give results close to the DES at almost twice the cost of the URANS method.

It is therefore advisable to adopt a triple strategy that begins with the evaluation of the URANS and DES for propeller flows and then compare the potential gains of SAS with the promised reduction in CPU time in comparison to the DES method.

The CFD results of any of the above method can be combined with a number of tools for the further exploitation of the pressure field in conjunction with far-field aeroacoustics methods.

Glasgow has experience with the FW-H method that is popular in the field of helicopter rotors. The method uses the CFD-generated unsteady pressure and based on the linearized acoustics equations, produces the acoustic signature of the propeller at distances far apart from the source of noise.

Thickness, loading and broadband noise sources could be resolved in the near-field and propagated further of the CFD domain with the FWH method. In addition, trailing edge noise should be resolved by a fine-mesh DES solution.

A second method that could be combined with the CFD results could lead to the prediction of the noise-level inside the cabin of an aircraft equipped with propellers. This would lead to an integrated simulation environment where the propeller performance and its acoustics could be studied and used for the far-field and cabin noise predictions at the same time.