CFD analysis of wing-propeller interaction on the NASA X-57 Maxwell aircraft wing

Due to global warming concerns, the Aviation industry is trying to reduce its carbon footprint. Electric propulsion (EP) is one way of doing this, where the power is obtained from electrical sources. The concept of distributed electric propulsion (DEP) is in the focus now. NASA’s X-57 Maxwell, a high winged, all-electric experimental aircraft, uses this concept. The present work aims at developing a CFD model (ANSYS Fluent) to evaluate aerodynamic performance of two configurations of NASA’s X-57 aircraft wing; (i) wing and nacelle (clean wing) and (ii) wing, nacelle and one electric propeller under cruise condition; and compare it with the results of wind tunnel experiment performed by NASA/Armstrong X-57 research program. Parameters like lift, drag and pressure coefficients (CL, CD, CP) are compared for both cases. A good match is observed for CL, CD and CP, thus validating the model. The unsteady RANS solver is very efficient in capturing the effects of propeller slipstream on the wing. After validation, this model is further used to simulate aerodynamic performance of a wing with multi-propeller (DEP) configuration.


Introduction
The air transport industry has been advancing, increasing its size twofold every twenty years.This led to high consumption of fuel, which in turn led to an increase in greenhouse gas emissions as well as noise.To limit its contribution to global warming, the aviation industry aims to reduce CO2 emissions in half by 2050, compared to 2005 levels [AIAA, 2021].This can be achieved by putting into practice innovative ways of propulsion, namely hybrid propulsion or all-electric propulsion.In hybrid propulsion, a combination of internal combustion engines and electric motors is used, while in allelectric propulsion, batteries are used as the main source of power.Conventional propulsion systems are complicated, require high temperature materials and special techniques for production, thus 2 proving to be quite expensive.Shifting the focus to unconventional propulsion systems has the potential to reduce fuel usage and toxic emissions, and also considerably reduce cost of production.Electric propulsion could quite possibly help in reducing fuel consumption by 90%, make the required power independent of flight conditions, and provide better reliability compared to conventional ICE propulsion system [Patterson et al., 2016].
Hybrid propulsion can be applied using 4 configurations: series, parallel, series/parallel, and turboelectric hybridization.It can be applied suitably to both large commercial aircrafts and smaller regional aircrafts.However, considering the weight of electrical equipment as well as fuel tanks, the operating empty weight of the aircraft could be higher.This could require a compromise between flight range and fuel saving.
All-electric propulsion has an impressive scope for development.It is a scalable technology that can easily be implemented in small as well as large scale aircrafts.The power density of batteries and energy efficiency of electric motors are developed and improved continuously, making this technology more viable.All-electric propulsion is estimated to reduce maintenance and fuel costs by 50% [Manuel Randon et al., 2021].
In conventional propulsion systems, the turbine needs to be coupled with the engine, which brings limitations in terms of operation.In electric propulsion, the propellers are decoupled from the EMs, which allows each propeller to be operated at optimal conditions.Even though EMs produce less thrust as compared to ICEs, due to their light weight and smaller size, several EMs can be utilized to produce the required thrust and improve performance.This concept of fitting several EMs on the aircraft is known as distributed electric propulsion (DEP).
DEP consists of a propulsion system that is closely integrated with the aircraft structure.It is a multi-propeller configuration where the propellers, driven by EMs, are distributed along the wing and/or fuselage (propulsive fuselage concept) [Kevin R Moor et al., 2018].This configuration is capable of producing high lift in a short time, which can reduce the distance required for takeoff.Numerous DEP aircrafts are under development for STOL and VTOL features.The NASA X-57 Maxwell, Aurora Flight Sciences XV-24 Lightning Strike, Joby Aviation S2, Lilium Jet, Airbus Vahan VTOL aircrafts are some examples [Kim Hyun et al., 2018].
Pavel Hospodář et al. (2019) applied DEP configuration to a modified wing of a general aviation 10-seater aircraft.The modification was that the wing area was reduced by half.They applied CFD analysis using a RANS OpenFoam solver and applied Spalart-Allmaras turbulence model to solve for compressible steady flow.For the clean half wing, the lift coefficient decreased; however, when DEP was applied, the half wing provided the same lift as the original wing.
The NASA X-57 Maxwell all electric experimental aircraft is being developed by NASA to employ the DEP configuration.It is a modification of the Tecnam P2006T aircraft, where the twin engines are replaced by two electric motors on the wing tips, and 6 smaller motors along the leading edge of each wing, and the wing area is reduced from 145 sq.ft. to 55.1 sq.ft.The wing-tip motors are cruise motors that help to reduce drag from wingtip vortices, while the six smaller motors provide high lift during takeoff and landing.The smaller propellers are foldable once in cruise condition, for further drag reduction and optimized energy consumption.This concept provides higher lift, dynamic pressure at low speeds, propulsive efficiency, and lower drag [NASA 2022].
NASA conducted wind tunnel tests for the X-57 Maxwell using the Lockheed Martin Low Speed Wind tunnel (LSWT), as part of the NASA/Armstrong X-57 Research programme.They conducted tests for a clean wing case and a wing-tip propeller mounted wing case.A scaled down model was used for the tests, and parameters like velocity wake, lift, drag, and pressure coefficients were observed.In June 2019, NASA conducted the AIAA Workshop for Integrated Propeller Prediction, with the goal of making available a database that can be used by other researchers for validating their CFD models.The validated CFD models can be used for accurate prediction of wing-propeller interactions, and can possibly reduce cost of development.This promotes the use of CFD analysis for the NASA X-57 Maxwell, thus accelerating the development of the aircraft prototype [NASA, 2019].The database consists of results from tests carried out for Mach 0.04, 0.08, 0.11, angles of attack from -10 o to +20 o , for CT 0.0 (propeller off) and 0.04 to 0.4 (propeller on), and aileron deflection from -45 o to +45 o .The wing model was digitally scanned to provide a 3D CAD model, and some meshes were also made available for CFD use.
The aim of the present work is to develop a RANS based CFD model to assess the performance of the NASA X-57 Maxwell aircraft wing, by referring the data from WIPP workshop for validation of two cases: (i) clean wing and (ii) wing-tip mounted propeller wing.This validated model is then used to analyze the aerodynamic performance of a DEP configuration on the wing.The original wing-tip propeller geometry, that was obtained from the workshop, is modified to have 6 smaller propellers distributed along the leading edge of the wing.The emphasis on developing a RANS solver for this work is to reduce computational time that is usually quite high for solvers like LES and DES, while maintaining the accuracy of results, and making a humble contribution to the development of the NASA X-57 Maxwell aircraft.

Clean wing case
The geometry and mesh were obtained from WIPP.A steady state, pressure-based RANS (Reynolds-Averaged Navier Stokes) solver was set up using ANSYS Fluent.The k-ω SST turbulence model with air as ideal gas, coupled scheme and second order discretization was used.Boundary conditions were based on parameters from the wind tunnel experiment: Inlet velocity 27.2 m/s (Mach 0.08), 15o temperature, no slip wall and pressure inlet.Reference values for wing surface area and mean aerodynamic chord were taken as 0.43587 m2 and 0.25781 m respectively.Residuals were set to 10-6, and calculations were run for AOA 0, 5, 7, 15 and 17 degrees, until convergence.

Wing-tip mounted propeller case
The geometry available from WIPP was modified using CATIA to include a rotating domain around the propeller.Initially, calculations were run for only the isolated propeller, for validation of the required rotational speed to produce thrust specified by WIPP.Fig. 1 shows the plot for thrust vs rotational speed.The rotational speed of 640 rad/s was considered for further calculations, as it provided a thrust of 80 N, which corresponded with the experimental results.A hybrid, unstructured mesh was created using ANSYS Meshing, mainly consisting of tetrahedrons with prismatic inflation layers near the wing and propeller blades' surface.A mesh dependence study that was done for meshes with 1.5, 6, 12 and 24 million elements.Fig. 2 shows the plot for drag coefficient values for these meshes.A difference of less than 5% was observed between the 12 and 24 million elements meshes.Thus, the 12 M elements mesh was optimized and used for further calculations.The solver setup in ANSYS Fluent was the same as the clean wing case, with additional settings for the propeller: frame motion in cell zone condition, where speed and axis of rotation were specified.

Multi-propeller (DEP) case
The wing-tip mounted propeller geometry was modified in CATIA by adding 6 smaller propellers along the leading edge of the wing, as observed in Fig. 3.It was necessary to define flight conditions before setting up the calculations.Using the same method as for the wing-tip propeller, isolated propeller calculations were carried out for the smaller propeller.A thrust vs rotational speed plot was obtained, as shown in Fig. 4. Equal power distribution among all propellers was assumed.The thrust ratio was assumed as corresponding to the power ratio of the propellers.The NASA X-57 Maxwell datasheet was referred for power values of the propellers, while the thrust values were taken from the isolated propeller calculations.The thrust ratio was closest to the power ratio when the rotational speed of the smaller propeller was considered to be 1680 rad/s.There were concerns regarding the availability of an electric motor that provides this kind of speed; however, the model used for analysis is a 40.5% scaled down version.When the model is scaled up, the required rotational speed is not as high, and finding an appropriate electric motor is manageable.From the datasheet, the power for cruise motor is 60 kW and that for the smaller propeller is 10.5 kW.
The power ratio was taken as: The thrust for the cruise motor is 74.2 N for 620 rad/s.From Fig. 3.10, if the thrust for smaller propeller is considered to be 11.5481N, the thrust ratio is: Once the flight conditions were defined, a mesh was created based on the mesh specifications from the wing-tip mounted propeller case, and steady state calculations were performed.

Clean wing case
Lift, drag and pressure coefficients from the CFD calculations were compared with experimental data.Fig. 5 shows the plot for comparison of lift curves.As observed, there is a good match in the values.A slight difference is observed for higher AOA, which is expected due to over-prediction of flow separation by the turbulence model.Fig. 6 shows a schematic of the wing, with locations at which pressure coefficients were measured for the wind tunnel tests.Figures 7 and 8 show the plot for comparison of pressure coefficients of CFD calculations with experimental data, at locations 44.386 and 60.955 inches.As observed, there is good agreement of the values, and thus the clean wing case is considered to be validated.

Wing-tip mounted propeller case
Fig. 9 shows the comparison of drag polars values obtained from CFD calculations to those from the experimental data.There was a mismatch for values at higher AOA.Also, there was a mismatch for pressure coefficient values at 60.955 in, at the leading edge; i.e., in the wake of the propeller.After trying out different turbulence models and settings (GEKO model, curvature correction, etc.), it was observed that steady state calculations were not sufficient to provide accurate results.Thus, a transient (unsteady, time-dependent) solver was set up, with k-ω SST turbulence model.The settings from steady state solver were applied, with the only difference being Mesh Motion instead of Frame Motion for the propeller.Calculations were run for 5000 timesteps, with a timestep size of 10 -4 s, and 40 iterations per timestep, having an angular increment of the propeller position of 0.036 o per timestep.Initially, a timestep size of 10 -3 was taken; however, it proved to be quite large, as the results were similar to the steady state solver.A timestep sensitivity study was not performed.The transient solver proved to be successful in providing accurate results, for the drag polars as well as the pressure coefficient plots.Fig. 10 shows pressure coefficient plots at 60.955 in.As observed, there is an improvement in results for transient calculations from steady state calculations.This validates the wing-tip mounted propeller case, and the model is now calibrated for DEP configuration analysis.

Multi-propeller (DEP) case
As there was no experimental data available from WIPP for this case, the results were compared with steady state results of the wing-tip mounted propeller case.Figures 11 and 12 show the comparison of drag polars, including and excluding the propeller respectively.For the propeller included case, there is significant reduction in drag, which was as expected.For the propeller excluded case, there is a slight increase in lift, and also an increase in drag.This is due to the addition of nacelles on the wing which reduce the smoothness of the wing surface.The results from CFD calculations for the multi-propeller wing case also show agreement with the findings of Pavel Hospodář.Fig. 13 shows the comparison of lift curves for multi-propeller case with wing-tip mounted propeller case.There is definitely an improvement in aerodynamic performance due to the application of DEP to the wing.Fig. 14 shows a comparison of lift curves for a normal wing, a wing with half wing area and DEP on the half wing, from the research done by Pavel Hospodář.It is observed that the lift decreases for the half wing, however, when DEP is applied, it shows performance similar to the original wing.Thus, the preliminary results are satisfactory.The smaller propeller blades in this geometry were scaled down from the larger propellers.However, those were not the optimal blades used in the NASA X-57 Maxwell.A new propeller geometry was taken from the VSP3 model available on the NASA X-57 website, and the geometry was modified.Fig. 15 shows the new geometry.

Conclusion
The present work was dedicated towards applying the RANS based CFD solver of ANSYS Fluent to analyze wing-propeller interactions for the NASA X-57 Maxwell aircraft.Clean wing and wing-tip mounted propeller cases were validated using data from wind tunnel tests provided by WIPP.The findings showed that the steady RANS solution was unable to capture the wing-propeller interaction.However, the unsteady RANS solver was able to capture it accurately, and showed good agreement with the experimental results, especially for the CP plots.This solver was then used to simulate the aerodynamic performance of the DEP configuration of the wing.Preliminary calculations showed good agreement with findings from other researchers.In this way, this study is able to contribute to the development of the aircraft.It also shows that a RANS based solver is capable of providing good results and can save a lot of computational time for researchers, compared to LES, DES based solvers.