CFD Comparison of Selected Distributed Propulsion Configurations for an Inverted-Joined-Wing Airplane

The design freedom associated with distributed propulsion systems allows for the consideration of various types, numbers, and sizes of propulsors. In the present work, two types of propulsors were considered:

• Foldable propellers of diameter 240 mm, and

• Electric Ducted Fans (EDFs) of internal diameter 76 mm.

Both options require 10 propellers to deliver sufficient thrust.

Five concepts, presented in Figure 2, were selected for CFD analyses. These comprise configurations with a row of small propellers located either in front of (b) or behind (c) the aft wing; a hybrid concept (d) using the original pusher propeller supplemented by an additional row of three small propellers; and two configurations using ducted fans, placed on the upper surface (e) or behind the aft wing (f). The concepts and their features are summarized in Table 1.

Fig. 2.

Selected configurations of distributed propulsion for MOSUPS.

Propellers were considered because the incidence angles of both wings were selected to achieve economic airspeed with a horizontal fuselage attitude. As a result, the MOSUPS requires neither significant rotation for take-off nor significant flare before landing (see Figure 1), meaning blade tips would maintain sufficient clearance from the runway. Furthermore, the risk of ground contact during landing can be reduced further, as the engines can be set to idle and the propellers folded. However, both the Prop 1 and Prop 2 configurations may present difficulties during a crosswind landing abort, as propeller blade tips could strike the runway if pitch, yaw, and bank angles are significant. For this reason, abort landing configurations were also investigated, in which the four outboard motors are shut down and their propellers folded, leaving only the six inboard engines available for the emergency maneuver. The Hybrid configuration did not require such a variant, as the outboard engines are absent, replaced by the central pusher.

Fig. 3.

Configurations with pulling and pushing propellers during crosswind landing.

Finally, to prevent propeller-runway contact in the Prop 2 configuration, additional vertical surfaces were added below the aft wingtips.

The CFD simulation was conducted using the finite volume method to solve the Navier-Stokes equations, which describe viscous fluid motion. They are derived by applying Newton’s second law with the assumption that stress in a fluid is the sum of viscous and pressure terms.

Continuity equation:

1 ∂ρ∂t+∇·(ρu→)=Sm {{\partial \rho } \over {\partial t}} + \nabla \cdot(\rho \vec u) = {S_m}

Momentum conservation equations:

2 ∂∂t(ρu→)+∇·(ρu→u→)=−∇p+∇·(=τ)+ρg→+F→ {\partial \over {\partial t}}(\rho \vec u) + \nabla \cdot(\rho \vec u\vec u) = - \nabla p + \nabla \cdot(\overline {\bar \tau } ) + \rho \vec g + \vec F

where p is static pressure, τ = \mathop \tau \limits^ = is the stress tensor, and ρ g → \rho \vec g and F → \vec F are the gravitational and external body forces. The stress tensor is defined as:

3 τ==μ[ (∇u→+∇u→T)−23∇·u→I ] \overline {\bar \tau } = \mu \left[ {\left( {\nabla \vec u + \nabla {{\vec u}^T}} \right) - {2 \over 3}\nabla \cdot\vec uI} \right]

where μ is the molecular viscosity, I is the unit tensor.

A commercial software package, ANSYS Fluent, was chosen for this task. Due to the complexity of the aircraft geometry, a structured mesh was not feasible; consequently, an unstructured tetrahedral mesh was created. ICEM CFD, a tool within the ANSYS suite, was used for mesh generation. This choice was dictated by ICEM’s powerful and robust prismatic mesh generation capability. Prismatic elements are extruded from the walls into the tetrahedral mesh, allowing even complex geometries to be enclosed by a well-defined boundary layer mesh. During this process, border elements may become compressed, so the tetrahedral elements are continuously smoothed and redefined as necessary.

In the calculations described, the Menter k-ω SST turbulence model was used [18]. This model requires a mesh satisfying a y⁺ criterion of 1, so the boundary layer mesh parameters were set accordingly. These settings allowed the solver to capture laminar separation bubble regions on the wing surface, which was one of the key objectives of this study. The resulting mesh is shown in Figure 5. To simulate the influence of the propeller wake, a simplified approach employing actuator discs in place of propellers was used. The pressure jump was determined by the required – or assumed – thrust value, and was held constant along the propeller radius. This approach is sufficiently simplified to enable calculation of multiple cases without detailed knowledge of the propeller geometry (rotational speed, airfoil sections, chord distribution, and blade pitch), thus avoiding the typical complications and extensive requirements of rotating propeller simulation.

Fig. 4.

Pushing variant with additional fins.

Fig. 5.

Mesh used for calculations.

Computations were performed across a range of angles of attack, with thrust set for cruise at an airspeed of 25 m/s. The quantitative and qualitative analyses allowed the effect of distributed propulsion on the MOSUPS to be assessed for each configuration considered. Figure 6 shows regions of reversed flow on the MOSUPS surface in the cruise configuration. As can be seen, none of the configurations resulted in complete elimination of the reversed flow region on the aft wing. For the Prop 1, Prop 2, and Hybrid configurations, the reversed flow region appears thinner in the chordwise direction compared with the baseline. For the EDF 1 and EDF 2 configurations, the reversed flow was eliminated entirely ahead of the fans but widened on either side of them. As a result, the excess drag associated with this phenomenon was not eliminated. Additionally, the engine nacelles increased the total drag of each configuration (Fig. 7).

Fig. 6.

Reverse flow regions on the top surface of MOSUPS in cruising configuration.

Fig. 7.

Effect of engines nacelle on aerodynamic efficiency CL/CD and lift coefficient CL in configurations Prop-1 and Prop-2, in comparison with baseline case without the thrust effect.

The main objective of the present study was to determine the effect of various distributed propulsion concepts on MOSUPS trim drag. Figure 8 presents the influence of the considered configurations on the pitching moment coefficient with no control surface deflection. It should be noted that the baseline configuration achieves trim at a relatively low angle of attack of -6°. Almost all distributed propulsion concepts improved this metric, shifting the zero-pitching-moment angle to around 2° in the EDF 1 case. The Hybrid configuration alone did not yield any significant improvement and is therefore not included in Figure 8.

Fig. 8.

Pitching moment for analyzed concepts. The influence of modifying the propulsion axis position and propeller selection reveals very large effects.

A corresponding reduction in trim drag was therefore expected. However, in the Prop 1 configuration, the adverse effect of thrust axis position was replaced by an adverse effect of increased aft wing drag and lift due to elevated dynamic pressure (Fig. 9), and the shift in zero-pitching-moment angle proved small. In the Prop 2 configuration, by contrast, the expected pitching moment improvement was achieved. Consequently, the balance of forces and moments occurred closer to the maximum L/D ratio without any elevator deflection (Fig. 10). The net effect of the Prop 1 configuration was unsatisfactory, yielding only L/D ≈ 7 against an available L/D_max ≈ 12. The Prop 2 configuration performed somewhat better, achieving L/D ≈ 9.5. However, Prop 2 carries the greatest risk in crosswind landing conditions and was therefore also analyzed with additional fins below the aft wingtips (Fig. 4). Figure 11 shows that the addition of fins reduces L/D_max slightly, and a trimmed value exceeding 9 should not be expected.

Fig. 9.

Difference in pressure distributions on the aft wing between tractor and pusher distributed propulsion configurations (AoA=0°).

Fig. 10.

Thanks to increased pitching moment, trim is achieved at higher angle of attack, which improves cruise efficiency. Vertical lines were added to indicate the trim position on the CL/CD plot.

Fig. 11.

The effect of skid installation below the aft wingtips.

A further means of improving safety was offered by the EDF configurations. The configuration with EDFs installed on the upper surface of the aft wing is optimal from this perspective. However, the associated increase in velocity also produces the strongest adverse effect on aft wing drag in this configuration (Fig. 12). As a result, the zero-pitching-moment angle was shifted to a favorable value, but the available L/D_max was simultaneously reduced to 9.5. The EDF 1 configuration therefore offers no performance advantage over Prop 2.

Fig. 12.

The effect of increased velocity ahead of EDF in EDF 1 and EDF 2 configurations.

Fig. 13.

Due to increased pitching moment, trim is achieved at higher angle of attack, which improves cruise efficiency. This is especially significant in EDF cases. Vertical lines were added to indicate the trim position on the CL/CD plot.

For this reason, the configuration with EDFs installed behind the aft wing was also investigated. This configuration is considerably safer than Prop 2 and offers the best performance of all cases analyzed. The zero-pitching-moment angle with no elevator deflection was shifted to 0°, while a high L/D_max ≈ 12.5 is maintained at AoA = 2°. Consequently, a value of L/D ≈ 12 can be expected in the cruise configuration.

The present paper describes an analysis of several distributed propulsion concepts as an attempt to address the pitching moment deficiency of an inverted joined wing aircraft. Additionally, local aerodynamic features of the aft wing – specifically, regions of reversed flow – were investigated. The study confirmed the potential of distributed propulsion as an extension of aircraft design freedom. Depending on the configuration, different degrees of improvement in flight performance were achieved. The EDF configurations proved the most promising, owing to their stronger influence on the flow field around the wing. Conventional propellers mounted at a distance from the wing had a comparatively modest effect on aft wing aerodynamics; their advantage over the baseline in the present study was attributable primarily to the change in thrust axis position rather than aerodynamic interaction with the wing.