Shapley Additive Explanations of Multi-Geometrical Variable Coupling Effect in Transonic Compressor

Optimization algorithms in the compressor detailed design stage generate big data of geometries and corresponding performances, but these data are often not exploited efficiently to unveil hidden compressor design guidance. In this work, the Shapley Additive Explanations (SHAP) method from game theory is proposed as an efficient methodology to extract design guidelines from databases. A database was generated when optimizing the blade features (sweep, lean, end-bend) of Rotor 37. Based on this, a neural network is trained to predict compressor efficiency. The SHAP method is then applied to explain the neural network behavior, which provides information on the sensitivity of single geometrical variables and the coupling effect between multiple geometrical variables. Results show that the near-tip sweep and mid-span lean angles are most influential on efficiency. Within the same group of variables, the adjacent variables tend to present strong positive coupling effects on efficiency. Among different groups, evident coupling effects are observed between sweep and lean and between lean and end-bend, but the coupling effect between sweep and end-bend is negligible. Flow mechanisms behind the coupling effects are discussed. For near-tip lean angles L3 and L4, the positive coupling effect is due to the change of the passage shock. For near-tip lean angle L4 and sweep angle S4, the change of detached shock leads to a negative coupling effect. The proposed data mining method based on the neural network and SHAP is promising and transferable to other turbomachinery optimization databases in the future.

Maximizing the performance of pump inducers using CFD-based multi-objective optimization

Pump inducers are usually employed within a limited flow rate range since the performance is known to drop out significantly far from their design point. Therefore, finding an optimal geometry that ensures efficient operation for a relatively wide range of flow rates is challenging. The present study tackles this problem using multi-objective optimization to identify optimal inducer configurations, delivering high performance for a wide flow range. 3D RANS single-phase turbulent simulations were performed using the $$k-\omega$$ k - ω turbulence model. The optimization was done by employing the Non-dominated Sorting Genetic Algorithm (NSGA-II) coupled with computational fluid dynamics (CFD). An established in-house flow optimization library (OPAL++) was used to automatically control the numerical simulations. The objective is to optimize the inducer geometrical parameters to simultaneously maximize the efficiency and pressure head curves, considering different flow rates, i.e., 80% (part-load), 100% (nominal), and 150% (overload) of the optimal flow rate for the considered pump. The optimization involves 8 most relevant design parameters, i.e., the axial blade length, blade sweep angle, blade pitch, hub taper angle, tip clearance gap, blade thickness at the hub, blade thickness at the tip, and the number of blades. A total of 5178 simulations over 37 generations have been needed to get a Pareto front containing 5 optimal configurations. This article discusses quantitatively the influence of each geometrical parameter on flow behavior and inducer performance. The results reveal in general that blade length, blade sweep angle, tip clearance gap, and blade thickness should be kept low for the considered application; inducers with high hub taper angles and 3 blades lead to optimal performance.

Dynamic Stall Characteristics of Pitching Swept Finite-Aspect-Ratio Wings

An experimental investigation regarding the dynamic stall of various swept wing models with pitching motion was performed to analyze the effect of sweep on the dynamic stall. The experiments were performed on a wing with a NACA0012 airfoil section with an aspect ratio of AR = 4. The experimental study was conducted for chord-based Reynolds number Rec =2×105 and freestream Mach number Ma=0.1. First, a ‘particle image velocimetry’ (PIV) experiment was performed on the wing with three sweep angles, Λ=0o, 15o, and 30o, to obtain the flow structure at several wing spans. The results obtained at a reduced frequency showed that a laminar separation bubble forms at the leading edge of the wing during upward motion. As the upward pitching motion continues, a separation burst occurs and shifts towards the wing trailing edge. As the wing starts to pitch downward, the growing dynamic stall vortex (DSV) vortex sheds from the wing’s trailing edge. With the increasing sweep angle of the wing, the stall angle is delayed during the dynamic motion of the wing, and the presence of DSV shifts toward the wingtip. During the second stage, a ‘turbo pressure-sensitive paint’ (PSP) technique was deployed to obtain the phase average of the surface pressure patterns of the DSV at a reduced frequency, k=0.1. The phase average of pressure shows a distinct pressure map for two sweep angles, Λ=0o, 30o, and demonstrates a similar trend to that presented in the published computational studies and the experimental data obtained from the current PIV campaign.

Influence of Various Stator Parameters on the Open-Water Performance of Pump-Jet Propulsion

In order to improve the hydrodynamic performance of pump-jet propulsion (PJP) when matching stator with the rotor, the RANS method with SST k-ω turbulence model is employed to study the influence of six kinds of stator parameters, which are classified into three groups, i.e., stator solidity, stator angles and rotor–stator spacing (S). Results show that the stator solidity involves the blade number (Ns) and chord length (L), has an obvious acceleration effect at and after stator, and produces a higher thrust and torque with a slight efficiency change. Further comparing Ns and L results, we find greater distinctions between the two cases when stator solidity is greatly adjusted. Three stator angles, i.e., stagger angle (α), lean angle (γ), and sweep angle (β), are studied. The α has the biggest effect on the thrust, torque, and efficiency; meanwhile, it shifts the advance number that corresponds to maximum efficiency. The effect of γ is similar to α, but its influence is far less than α. However, there is little difference between various β cases except for off-design conditions, where the efficiency drops dramatically as β increases. The S has a slight effect on PJP performance. Even though S decreases 34% relative to the original PJP, the rotor thrust and torque increase by less than 1%. In addition, we compare torque balance locations under various parameters, and each component force is analyzed in detail to explain the reason for performance variation. The present work is conducive to future optimization in PJP design.

The KhAI-90 Light Civil Turboprop Airplane Pilot Project

The pilot project of new light civil turboprop aircraft, called the KhAI-90, featuring a cruising speed of 350km/h, payload of 600 kg at 500 km range, and equipped with two turboprop Rolls-Royce 250-B17F engines each with power of 420 hp (alternatively, two AI-450C engines each with power of 450 hp may be installed) is presented herein. Based on the developed technical task, the concept for creating the KhAI-90 new competitive light civil aircraft, and the analysis of prototypes’ aircraft parameters and characteristics, the main tactical and technical requirements are assigned. The take-off weight of the new aircraft is determined in three approximations at the preliminary design stage of light civil turboprop aircraft, using the iterative software “CLA-TOW”, studying the influence of the wing geometric parameters and lift devices on aerodynamic performance, the power-to-weight ratio and the airplane weight parameters. The following parameters are calculated for the design: minimum take-off weight WTO min = 3,600 kg, optimal wing loading p 0 opt = 130 daN/m2, optimal aspect ratio 9.6, taper ratio 2.25, sweep angle at leading edge 3 degrees, airfoil relative thickness 10.6%. A general view and three-dimensional parametric models of the master-geometry and passenger cabin space distribution are constructed for the KhAI-90 by means of the SIEMENS NX computer integrated system. More broadly, this pilot project has also demonstrated the viability of the method we developed and previously reported for determining light civil turboprop airplane parameters.

The Impact of Sweep Angle on Stepped Planing Hull Performance

In this study the impact of sweep angle on stepped hull resistance, running attitude, and dynamic stability is investigated for a range of planing speeds from ventilation inception (𝐹𝛻≈2) to high planing speeds (𝐹𝛻≈7) using RANS CFD. Potential performance benefits of the step are isolated for three speeds and two displacements using fixed trim simulations. Differences in running attitude and dynamic stability are investigated using free running simulations at the highest speed for a range of LCG locations. Finally, any differences in ventilation inception and performance at low speeds are investigated using fixed trim and heave simulations. The study shows that swept forward steps do not necessarily ventilate earlier than other step designs but do provide resistance reductions at 𝐹𝛻<5 compared to swept aft and unstepped designs. However, at 𝐹𝛻>5, swept forward steps demonstrate significant resistance increases compared to unswept and swept aft steps. At high speeds, swept aft steps provide improved dynamic stability compared to other step designs without a resistance penalty when compared to unswept steps.

Coverage Strategy for Target Location in Marine Environments Using Fixed-Wing UAVs

In this paper, we propose a coverage method for the search of lost target or debris on the ocean surface. The OSCAR data set is used to determine the marine currents and the differential evolution genetic filter is used to optimize the sweep direction of the lawnmower coverage and get the sweep angle for the maximum probability of containment. The position of the target is determined by a particle filter, where the particles are moved by the ocean currents and the final probabilistic distribution is obtained by fitting the particle positions to a Gaussian probability distribution. The differential evolution algorithm is then used to optimize the sweep direction that covers the highest probability of containment cells before the less probable ones. The algorithm is tested with a variety of parameters of the differential evolution algorithm and compared to other popular optimization algorithms.

CFD analysis of variable geometric angle winglets

Purpose Winglets play a major role in saving fuel costs because they reduce the lift-induced drag formed at the wingtips. The purpose of this paper is to obtain the best orientation of the winglet for the Office National d’Etudes et de Recherches Aérospatiales (ONERA) M6 wing at Mach number 0.84 in terms of lift to drag ratio. Design/methodology/approach A computational fluid dynamics analysis of the wing-winglet configuration based on the ONERA M6 airfoil on drag reduction for different attack angles at Mach 0.84 was performed using analysis of systems Fluent. First, the best values of cant and sweep angles in terms of aerodynamic performance were selected by performing simulations. The analysis included cant angle values of 30°, 40°, 45°, 55°, 60°, 70° and 75°, while for the sweep angles 35°, 45°, 55°, 65° and 75° angles were used. The aerodynamic performance was measured in terms of the obtained lift to drag ratios. Findings The results showed that slight alternations in the winglet configuration can improve aerodynamic performance for various attack angles. The best lift to drag ratio for the winglet was achieved at a cant angle of 30° and a sweep angle of 65°, which caused a 5.33% increase in the lift to drag ratio. The toe-out angle winglets as compared to the toe-in angles caused the lift to drag ratio to increase because of more attached flow at its surface. The maximum value of the lift to drag ratio was obtained with a toe-out angle (−5°) at an angle of attack 3° which was 2.53% greater than the zero-toed angle winglet. Originality/value This work is relatively unique because the cant, sweep and toe angles were analyzed altogether and led to a significant reduction in drag as compared to wing without winglet. The wing model was compared with the results provided by National Aeronautics and Space Administration so this validated the simulation for different wing-winglet configurations.

Stability Analysis and Augmentation Design of a Bionic Multi-Section Variable-Sweep-Wing UAV Based on the Centroid Self-Trim Compensation Morphing

In this study, a design scheme for a high-aspect-ratio bionic multi-section variable-sweep wing unmanned aerial vehicle (UAV) that utilizes the reverse coordinated change in the sweep angle of the inner and outer wing sections is proposed, which improves the aerodynamic performance and realizes the self-trim compensation of the wing’s centroid. According to the layout characteristics of this type of UAV, a reasonable distribution design of the wingspan ratio of the inner and outer sections is explored, to reduce the impact of aerodynamic center movement and moment of inertia change. The calculation and analysis results show that the coordinated variable-sweep scheme can significantly improve the influence of sweep angle change on the longitudinal static stability margin of UAVs with a high aspect ratio. The coordinated sweep angle change in the inner and outer wing sections can not only reduce the drag during high-speed flight, but also play a significant role in improving the performance of the aircraft at different stages in the mission profile. Appropriately increasing the wingspan proportion of the inner section can reduce the trim resistance of the V-tail, reduce the thrust of the engine, and increase the range and duration of the UAV. From the perspective of stability change, the multi-section variable-sweep wing UAV with a wingspan ratio of the inner and outer sections that is between 1.41 and 1.78 has better dynamic stability performance. Among them, the UAV with a wingspan ratio of the inner and outer sections that is equal to 1.41 has better longitudinal stability performance, while the UAV with a wingspan ratio of the inner and outer sections that is equal to 1.78 has better lateral/directional stability performance.

Triglobal infinite-wing shock-buffet study

This article uses triglobal stability analysis to address the question of shock-buffet unsteadiness, and associated modal dominance, on infinite wings at high Reynolds number, expanding upon recent biglobal work, aspiring to elucidate the flow phenomenon's origin and characteristics. Infinite wings are modelled by extruding an aerofoil to finite aspect ratios and imposing a periodic boundary condition without assumptions on spanwise homogeneity. Two distinct steady base flows, spanwise uniform and non-uniform, are analysed herein on straight and swept wings. Stability analysis of straight-wing uniform flow identifies both the oscillatory aerofoil mode, linked to the chordwise shock motion synchronised with a pulsation of its downstream shear layer, and several monotone (non-oscillatory), spatially periodic shock-distortion modes. Those monotone modes become outboard travelling on the swept wing with their respective frequencies and phase speeds correlated with the sweep angle. In the limiting case of very small wavenumbers approaching zero, the effect of sweep creates branches of outboard and inboard travelling modes. Overall, triglobal results for such quasi-three-dimensional base flows agree with previous biglobal studies. On the contrary, cellular patterns form in proper three-dimensional base flow on straight wings, and we present the first triglobal study of such an equilibrium solution to the governing equations. Spanwise-irregular modes are found to be sensitive to the chosen aspect ratio. Nonlinear time-marching simulations reveal the flow evolution and distinct events to confirm the insights gained through dominant modes from routine triglobal stability analysis.