CFD simulations have been performed for the cruise variant of the FTB2, using two differing fidelities – one based on steady RANS, with the propellers modelled by actuator disks, and one based on URANS, with the rotating blades modelled. For the former, the actuator disk databases were generated using 2D RANS applied to propeller blade sections from root to tip. Power-off simulations were also performed. The simulations, all at M0.2 covered a range of incidences, sideslips and Reynolds numbers, with the latter extending up to full scale. Generally, results from the more efficient steady RANS simulations agreed well with those from the less efficient, but in principle more accurate, URANS simulations, with significant differences only occurring at higher incidence angles.
The POLITE model was tested in the pressurised ONERA F1 facility, power-off. The full run matrix was completed, with data acquired from 1 Bar to 3.85 Bar; the highest tunnel pressure corresponds to approximately one half of full-scale Reynolds number. Wind-tunnel data was corrected for both strut tare and strut deformation effects, as planned. Unexpectedly, post-test analysis indicted that corrections were also necessary for strut blockage/interference. A correction process utilising historical data from a different test but with an almost identical strut was derived and successfully applied to the test data. The test data was delivered to ADS for input into the permit-to-fly activities for the FTB2.
The CFD was compared with the F1 power-off data and power-on data from a test in the RUAG facility undertaken as part of the CS2 ReLOAD project. Generally, there was good tie-up between the CFD and wind-tunnel test data for both airframe forces and moments and propeller loads – and this level of comparison extended to more sensitive quantities such as power increment.
Data fusion in the project came in two parts. Firstly, the two CFD fidelities were fused together, and secondly, the CFD was fused with the wind-tunnel data. Due to the complexities of having two sets of test data, one power-off at higher Reynolds number and one power-on at low Reynolds number, a bespoke process was required for the second data fusion activity. This was successfully generated, effectively giving a methodology for extrapolating sub-scale test data to full-scale Reynolds number using CFD.
The POLITE model could not be tested power-on in the F1 facility as it failed new EU regulations for pressure vessels – the propellers are driven by high-pressure air. To prepare the model for a future power-on test, should it be needed, appropriate parts of the POLITE model were re-manufactured in a material compliant with the EU regulations. The refurbished model then underwent a shakedown test in the Filton LSWT, to check operation of the power train and the propeller rotation rate regulation. This test was successful.
A new CFD methodology has been demonstrated, which is designed to significantly reduce non-physical (mesh) effects when vortices are convected through the flowfield. The methodology uses a combined Eulerian/Lagrangian approach. Further developments within the project have allowed this to be applied to a propeller example, showing the potential of the methodology to achieve improved accuracy at lower cost for configurations such as turboprops.
The outputs from PERTURB have already appeared in two papers at an international conference and two more are planned. In addition, these outputs have positively influenced other CS2 projects being performed at one of the beneficiaries. The results will be used to try and generate new business at one beneficiary, potentially in the area of distributed propulsion – and the other beneficiary is already using the results to propose follow-on research work in the propeller modelling area to be funded by national government.
