Aerodynamic, Aeroelastic and Aeroacoustic Ducted Propeller Optimisation for Regional Air Mobility

Methods

In contrast to CFD simulations of the ducted propeller conducted in previous reporting periods two modifications of the flow domain were made to strengthen numerical stability and reduce computational costs. First, the spinner was replaced by a trumpet-like type of contour to avoid stagnation point instabilities which led to unphysical high turbulence production in past simulations. Second, the inflow domain was shortened to reduce computational effort and avoid axial dampening of the applied flow distortion. The new setup was compared to older meshes by means of steady state simulations. The modifications of the flow domain show no influence on the overall performance of the ducted propeller. To determine the required step size for the subsequent distortion simulations, a study with three different time step sizes was performed. The TRACE User Guide suggests a time step of 420-700 steps per revolution as a rough estimate for the given application. As a consequence, simulations with 360, 720 and 1440 steps per revolution at two operating points were performed. One operating point is close to the stall line while the other is associated with smaller indicendes and less challenging flow conditions. The largest time step (360 steps per revolution) was found to be too coarse. For the less throttled operating points, differences between the mid-sized and fine time step were neglectable. However, differences in convergence behavior between the two smaller time steps for the operating point near the stability limit were observed. Therefore, the smallest time step was selected as the final time step for the remaining simulations. Test calculations were conducted with a total simulation time equal to ten revolutions. However, global performance parameters did not change significantly after five revolutions. The remaining simulations were limited to five revolutions for this reason. Five operating points without distortion using URANS were calculated as a reference for the 3000 rpm speedline by varying static pressure at the outflow. The applied distortion amplitude corresponds to a DC60 coefficient of ~0,5 which is considered a rather high distortion amplitude. The amplitude was kept constant for all simulations with inlet distortion. To capture possible performance losses when expanding the distortion sector, four angle segments, each of different size (15, 30, 60 and 120 Degrees), were chosen. Five operating points on the 3000 rpm speedline corresponding to the undistorted reference were calculated for every case. Simulations were initialized from steady state calculations of the particular combination of sector angle and operating point in order to reduce the required computational time of unsteady calculations to a minimum.

Results

The generated datasets are still subject to ongoing evaluations. However, first indications show that the ducted propeller is less sensitive to total pressure distortions than conventional jet engine fans. The calculated speedlines with distortion sectors of different sizes show no relevant loss in total pressure ratio compared to the clean undistorted case. Operating points closer to the stall line exhibit a growing boundary layer compared to the less throttled operating points. However, there is no complete flow separation on the suction side of the rotor blade present when the blade passes through the distorted sector. This behavior can be observed for all investigated distortion sector angles.

Discussion

Comparisons with analytical models and previous studies give an explanatory approach. Whether it is likely to observe a complete flow separation from the blades is highly sensitive to rotational speed and the chord length of a blade. The flow does not react instantaneously to a distortion, but with a certain delay. Large time constants are achieved mainly by two design variables. First, a high rotational speed which reduces the time spend in the distorted sector. Second, a long chord length which leads to a large time difference between stall inception and complete flow separation. The low total pressure ratio is another reason for the compressor not showing typical surge cycles. The potential energy stored in terms of compressed air between the rotor and the nozzle is too low. Analytical models also show a small to almost zero impact on the total pressure loss for inlet distortions in case of very small speedline curvature. This behavior is also observed and confirmed by the current numerical studies in this project. Future studies should consider a change of blade geometry (e.g. shorter or longer blade chords) to get more insights into the sensitivity of the blade geometry on inlet distortion resilience for low pressure fans.

Additional Project Information

DFG classification: 404-04 Hydraulic and Turbo Engines and Piston Engines
Software: TRACE, Python3, Tecplot
Cluster: CLAIX

Publications

Liegert, R. Experimental and numerical investigation of a propulsor with low pressure ratio under the influence of distorted inflow, PhD thesis, not yet published