Design and examination of a turboprop inlet with scavenge duct

Gaojie ZHENG, Zhenlong WU, Huijun TAN, Yue ZHANG, Kun WANG

College of Energy and Power Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

KEYWORDS Design methord;Flow characteristics;Inlet performance;Scavenge duct;Turboprop aircraft

Abstract The inlet with scavenge duct is an important part of turboprop aircraft engine.This type of inlet normally has a complex shape, of which the design is challenging and directly affects the flow field quality of the engine entrance and thus the engine performance.In this paper, the parametric design method of a turboprop aircraft inlet with scavenge duct is established by extracting and controlling the transition law of the critical characteristic parameters.The inlet’s performance and internal flow characteristics are examined by wind-tunnel experiment and numerical simulation.The results indicate that a flow tendency of winding up on both sides is formed due to the induction of the inlet profile, as well as a vortex pair on the back side of the power output shaft.The vortex pair dominates the pressure distortion index on the Aerodynamic Interface Plane (AIP).In addition, with the increase of freestream angle of attack, the total-pressure recovery coefficient of the inlet increases gradually while the total pressure distortion index decreases slightly.On the basis of the experimental results under different working conditions, the parametric design method proposed in this paper is feasible.

1.Introduction

Turboprop engine plays an important role in the field of regional airliners and military transport due to its high propulsion efficiency, low fuel consumption and good economic performance.1–3However, turboprop aircraft may face a variety of complex and severe flight environments when performing tasks, one of which is foreign objects.Once foreign objects are swallowed into the inlet,the engine performance will probably be degraded and even broken down, and thus the flight safety is highly threatened.Therefore,a turboprop engine inlet should have a high tolerance to foreign objects.4–7In order to prevent foreign objects from being sucked into the inlet and entering the engine, a foreign object removal device should be designed, which is called scavenge duct.8

The design of inlet with scavenge duct is an important part of turboprop aircraft design, which directly affects the flow field quality, engine performance and exclusion efficiency.9,10A previous study shows that if the total pressure recovery coefficient of the inlet is reduced by 1%,the thrust will be reduced by 1.5%-2.0%.11The design difficulties regarding turboprop engine inlet with scavenge duct lie at least in these aspects:(A) the power output shaft should pass through the center of the engine duct, which requires its section to gradually transition from the ‘‘runway shape”at the entrance of the inlet to the ‘‘ring shape”at the section of the engine duct.(B) The scavenge duct and the inlet make up a complex threedimensional bifurcation flow channel.Thus, the transition between the engine duct and the scavenge duct at the bifurcation location is also a key point in the inlet design process.(C)The inlet is confined by various factors such as the nacelle and offset distance, and should be well compatible with the gearbox and the structure of the nacelle.Taking the TP400 engine of the Airbus A400M as an example, the propeller and the engine are not on the same axis, so the inlet needs to make room for the power output shaft and the gearbox, which greatly increases the difficulty in the design of the‘‘swan neck”inlet of the engine.12

The complex surface configurations of turboprop engine inlets with scavenge ducts produce complicated internal flow field characteristics, which have not been fully studied so far.13–15In 2005,Ruiz-Calavera et al.16summarized the work of optimization design of engine front end including inlet based on the nacelle-inlet model of the Airbus A400M military aircraft.In 2010, their team17conducted a comparative study on the total pressure distribution and swirl distortion in the inlet AIP under the conditions of propeller power-on and power-off through numerical simulation and wind tunnel test.The experimental performance of the inlet was primarily focused, but the internal flow field was not discussed in detail and the factors that affect the performance of the inlet were not analyzed.It was also pointed out that it is difficult to design the turboprop inlet with scavenge duct, but no specific parametric design method was given.

In this paper,the parametric design method of a turboprop aircraft inlet with scavenge duct is established by extracting and controlling the transition law of the critical characteristic parameters.Wind-tunnel experiment and numerical simulation are conducted to examine the performance and internal flow characteristics of the validated inlet.

2.Parametric design method

2.1.Determination of inlet and outlet areas

The overall layout of the typical turboprop inlet with scavenge duct is shown in Fig.1.Prior to the generation of the inlet profile, the areas of the inlet throat, engine-duct outlet and scavenge duct outlet need to be determined first.As the design conditions, namely the flight state (flight Mach number Ma0and altitude H) and the engine working state, are determined,the engine duct mass flow rate ˙m1, the engine duct outlet area A1(namely the engine entrance area) and Mach number Ma1can be determined.According to the scavenge ratio between the engine duct and scavenge duct (SCR, usually less than 30%11),the scavenge duct mass flow rate ˙m2can be determined by

In the conventional subsonic inlet,the total pressure recovery coefficient is higher,and taking S-duct inlet as an example,when the outlet Mach number is around 0.4,the total pressure recovery coefficient is generally above 0.9818.The scavenge duct in this paper is approximately a straight duct without turning, and the flow loss is very small, so the reference value of σ2is 0.99.According to the design experience of subsonic inlet, the Mach number at the outlet of the scavenge duct is recommended not to exceed 0.8 to prevent flow congestion which will affect the performance of inlet under high-speed extreme conditions.According to the throat contraction ratio r, the throat area Atcan be determined, At= rA0, and r is about 0.85.

2.2.Solution of rotation angle and line segment length between discrete points in lateral direction

The engine duct profile of the inlet is composed of J sections as shown in Fig.2.The sections from the inlet section to the outlet section are 1, 2,..., J.The specific design process is described below.

Fig.2 Illustration of section profiles and discrete points.

According to Section 2.1 the section lines at the inlet throat,the engine duct outlet and the scavenge duct outlet are determined and discretized,as shown in Fig.2.The discretized section lines are expressed in point sets.For example, the discretization of the inlet and outlet section profiles are respectively the point sequences of (xin,i,yin,i,zin,i) and(xout,i,yout,i,zout,i), where i is the point sequence number,i=1,2,...,I.In the same section,the angle θibetween the line segment of two adjacent points and the horizontal line can be calculated by

2.3.Solution of rotation angle and segment length between discrete points in streamwise direction

According to the rotation angle and the length between discrete points of the inlet and outlet section profiles calculated in Section 2.2, the increment of the rotation angle of adjacent points between the inlet and outlet is calculated as

In order to obtain the rotation angle and line segment length of the corresponding points of each section in the streamwise direction, the transition law is set in the form of exponential function in this paper, i.e., cjm=（jm/J)a,jm/J ?［0,1], as shown in Fig.3, where cjmis the angular rotation and length change rate at the jmsection.Then at the same section, the angle θjm,ibetween the connecting segment of i point and i + 1 point and the horizontal line is

2.4.Solution of geometric profile of each section

After obtaining the rotation angle at each discrete point and the length of the line segment between adjacent points in the section, the geometric profile of each section can be obtained as long as the coordinates of a point in each section are known,as presented as follows.

Assuming that the coordinates of the first point of any jmsection are （xjm,1,yjm,1,zjm,1), and the rotation angle at the ith point of the section is θjm,i, the coordinates of the ith discrete point of the jmsection can be calculated by

The coordinates of the first point of the section are already known,and the z coordinates in the flow direction are equidistantly distributed between the inlet and outlet.At this point,the coordinates of the whole profile have already been solved.According to the change law of the center line,the height of the geometric profile of each section is adjusted.The change law of the center line can be referred to in Refs.19–21.In this paper, the fifth-order polynomial is selected, as shown in Fig.4.The vertical ordinate of discrete point in the jmsection is adjusted as follows:

where h is the offset distance between the entrance and exit of the inlet, and djmis the growth rate at the jmsection.

Fig.3 Control curve of transition law.

Fig.4 Control curve of centerline change rate.

Finally,the sections are scaled and rotated according to the change law of each section area in the streamwise direction.

So far, the parametric design of the inlet surface has been realized by determining the two main parameters a and c.Parameter a is the change rate that controls the rotation angle and segment length between discrete points, which affects the variation degree of the inlet section.Parameter c is the control parameter for determining the governing equation of the center line.Considering various design schemes and experience,a=2.5 and c=0 are chosen,which are beneficial to the performance of the inlet.The scavenge duct profile is generated according to the same design process.Then the engine duct and the scavenge duct are intersected with a fillet surface transition to realize the design of the turboprop inlet with scavenge duct, as shown in Fig.5.The corner radius R of the scaled model in this paper is 50 mm.

3.Experimental and numerical methodology

3.1.Test model and experimental setup

The inlet profile based on the design parameters was used for the wind-tunnel test, and the design parameters are shown in Table 1.The inlet model with nacelle was further designed and fabricated, as shown in Fig.6.The experiments were all performed in the high-speed NH-1 wind tunnel of Nanjing University of Aeronautics and Astronautics.The test section of the wind tunnel is 600 mm in height and 600 mm in width,and the unit Reynolds number ranges from 1.2 × 107m-1to 1.77 × 107m-1.The available test Mach number range is 0.3–2.0 and the test time of a single run is typically 60 s.Considering the blockage of the wind tunnel, the model was reduced by 9 times.The geometric parameters of the scaled model are shown in Table 1.The blockage of the test section of the wind tunnel was 4.9% with the test model installed.The angle of attack of the model is positive counterclockwise.

Fig.5 Inlet profile generated by parametric design.

Table 1 Geometry parameters of inlet.

During the tests,a number of pressure measurement points,including the surface pressure taps and total pressure tubes,were placed to monitor the internal flow structures and obtain the inlet performance.As shown in Fig.7 and Table 2, eleven rows of static pressure taps distributed along the streamwise and lateral directions were flush-mounted on the nacelle and the internal surface of the turboprop inlet to acquire the static pressure.The total pressures were measured by the rakes assembled at the exit plane of inlet(x/L=0.77,L is the length from the leading edge of the model to the end).As shown in Fig.6, each rake contained four branches, which were placed at an interval of 45°.Every branch has four pressure tubes.Eight static pressure taps were also placed on the duct surface to obtain the average static pressure of the AIP, which is necessary for the calculation of the average Mach number MaAIPand circumferential distortion index DC60of the AIP.The pressure was sensed by the electronically-scanned pressure system (ESP-64HD, PSI Inc?) with an error of 0.05% at full scale, which embraces two types of modules of different measuring ranges, i.e., 5 psi and 10 psi (1 psi = 6894.76 Pa).Furthermore,only the pressure data of one side were extracted for the in-depth analysis, whereas the rest was used for validation and not elaborated hereinafter.The test model installed in the NH-1 wind tunnel is shown in Fig.8.

Two most common parameters,the total pressure recovery σ and the circumferential distortion index DC60are used to quantify the inlet performance at the AIP.The total pressure recovery σ is defined as

where p is static pressure, p0is freestream static pressure, ρ0is freestream density, and V0is freestream velocity.

Fig.6 Sketch of nacelle/inlet and pressure taps configuration.

Fig.7 Locations of surface pressure taps.

Table 2 Information of pressure taps.

Fig.8 Test model installed in NH-1 wind tunnel.

3.2.Numerical setup and validation

To obtain the detailed flow field features of the inlet,the numerical method is utilized to offer a comprehensive revelation of the three-dimensional internal flow structures.The commercial Computational Fluid Dynamics (CFD) software ANSYS FLUENT was adopted to perform the calculations, which has been widely used for internal flow simulations22–24.FLUENT applies the finite-volume method to solve the threedimensional Reynolds-averaged Navier-Stokes (N-S) equations.The turbulent viscosity is resolved by the Shear Stress Transport (SST) k-ω model, which is proposed by Menter25and shows great advantages in predicting complex flows with adverse pressure gradients.

The computational domain is schematically presented in Fig.9.To eliminate the boundary effects on the CFD results,a large computational domain, 50D in width and 65D in length, was used.The domain is filled with hexahedral grids,which is created by ANSYS ICEM CFD.The surface grid of the nacelle and the inlet with bypass duct is shown in Fig.10.The grids are smoothly refined in the near-wall regions with the y+kept near 1 for most near-wall grids to satisfy the turbulence model’s requirements.The contours of total pressure recovery coefficient distribution on the AIP are drawn by interpolation of pressure measurement points, and the interpolation method is the biharmonic spline interpolation which is a widely used interpolation method.26,27

Fig.9 Details of computational domain and boundary conditions.

Fig.10 Surface mesh of nacelle and nacelle/inlet.

In order to verify the rationality of the selected grid scale,three kinds of computational grids with different spatial resolutions, i.e., coarse, fine, and dense, are constructed, with the grid size of 6.7 million,8.8 million and 12 million,respectively,while ensuring that the overall topology of the grid and the height of the first layer of the near-wall grid remain unchanged.The wall static pressure obtained by experiment and simulation is compared under the condition of incoming Mach number Ma0=0.5,AIP Mach number MaAIP=0.425 and throughflow at the scavenge duct.The static pressure distributions on the streamwise inlet surface obtained by the three kinds of grid are slightly different,as shown in Fig.11.In this paper,the fine grid is used for all the following numerical simulations.In addition, the simulation results are in good agreement with the experimental results,suggesting that the present numerical method has an acceptable accuracy for calculating the flow field of the turboprop engine inlet.

4.Results and discussion

4.1.Performance parameters and surface flow patterns at AIP

Fig.12 shows the total pressure recovery coefficient and distortion index of the engine duct with different outlet Mach numbers under the condition of freestream Mach number of 0.5 and throughflow in the scavenge duct.Besides, the contours of total pressure coefficient distribution on the AIP under the experimental outlet Mach numbers of 0.364, 0.425 and 0.561 are accompanied in the figure, with the corresponding DC60being 0.119, 0.094 and 0.062, respectively, satisfying the general requirement of flow uniformity of DC60< 0.40 for common gas turbines.11According to the experimental results,the total pressure recovery coefficient decreases gradually with the increasing outlet Mach number, which is due to the increase of the flow rate inside the inlet and the consequent increase of the loss along the inlet.The total pressure recovery coefficient always remains above 0.966 in the whole range of AIP Mach number of interest, suggesting that the designed intake port has a satisfactory overall performance.In addition,according to the contours of σ shown in the figure, with the increase of outlet Mach number, the low total pressure area near the outer wall of the inlet at θ=0°and the two low pressure areas adjacent to the power output shaft at θ=135°and θ = 225° all gradually increase, which is consistent with the above result of the decrease of the total pressure recovery coefficient.It is noteworthy that with the gradual increase of the AIP Mach number, the high total pressure area in the lower half of the section increases slightly, indicating that the flow in the lower half of the channel will be smoother as the blockage degree of the engine duct gradually decreases.

Fig.11 Experimental and numerical surface pressure distributions on inlet with different grids.

A comparison between the numerical and experimental contours of the total pressure recovery coefficient on the AIP is shown in Fig.13 for MaAIP= 0.425, in which the dashed circles denote the experimental measuring borders.Similar to the experimental results discussed above, one low total pressure region located on the outer wall as well as two near the power output shaft are also clearly seen in the numerical results.In general, the numerical results are consistent with the experimental results.In addition, similar good agreement is also reached between the experimental and CFD results qualitatively at MaAIP=0.385.In order to quantify the difference between experimental results and simulation results, the experimental results of 32 total pressure taps are compared with simulation at MaAIP=0.425,as shown in Fig.14,which can be seen that the current simulation has reached the general accuracy requirement.

Fig.12 Relations of inlet performance parameters with AIP Mach number.

4.2.Three-dimensional flow structures of inlet

The distribution of total pressure recovery coefficient and secondary streamline at the different sections along the streamwise direction are shown in Fig.15.The total pressure recovery coefficient distribution in the scavenge duct basically does not change along the streamwise direction in the throughflow state, while the distribution of the total pressure coefficient in the engine duct changes slightly along the streamwise direction.Three vortex pairs are generated in the section of Rin 1, of which Vortex pair 2 is the dominant one, while Vortex pair 1 gradually decays till it disappears finally along the streamwise direction.

Fig.14 Comparison of experimental and numerical results of σ at MaAIP = 0.425.

Fig.13 Contours of σ distribution on AIP: Experimental result (left) and CFD result (right).

Fig.15 Total pressure recovery coefficient distributions on different sections.

To demonstrate the generation mechanism of these three vortex pairs, the secondary streamline distribution in various cross-sections along the streamwise direction and the Q-criteria based vortex structure are shown in Fig.16 and Fig.17,respectively.Vortex pair 2 is the largest vortex in these three vortex pairs, and its formation process is related to the flow around the power output shaft.When the air is ingested into the inlet, it is divided into the scavenge duct and the engine duct under the guidance of the inlet profile.Due to the existence of the power output shaft, the airflow entering the engine duct will bypass the shaft at a large angle of attack and meet at its back.According to the high angle of attack flow prediction28–30, a stable vortex pair will be formed on the rear side after the airflow bypasses the shaft.Besides the influence of the power output shaft, the induction of the inlet profile is also an important reason for the formation of Vortex pair 2.The cross-section of the inlet gradually transitions from the ‘‘runway shape”at the entrance to the ‘‘circular shape”of the engine duct.A trend of rolling-up on both sides of the engine duct is gradually developed for the internal airflow under the guidance of the inlet profile.When the developing rolling flow reaches the power output shaft, the two separate flows meet at the top side, forming a transverse secondary flow, i.e., Vortex pair 2, in the cross-section.

Fig.16 Total pressure recovery coefficient distributions in various cross-sections along streamwise direction.

Fig.17 Illustration of vortex pairs in inlet based on Q-criterion vortex intensity.

For Vortex pair 1 near the wall with θ = 180°, the reason for its formation lies in the restraint by the annular channel between the inlet and the power output shaft, where the two strands of airflow encounter collisions near the corner region,as shown in Fig.16.As a consequence, the boundary layer is separated,forming Vortex pair 1,and a saddle point is formed in the cross-section, which has been shown in Fig.15.Later,the flow continues to evolve gradually along the streamwise direction.The momentum of the two meeting flows exchanges continuously and Vortex pair 1 also dissipates and disappears finally.Among the three vortex pairs,Vortex pair 2 has the largest intensity scale and the lowest total pressure in the vortex, which is dominantly responsible for the total pressure distortion index at the AIP.The other two vortex pairs have smaller scales and thus less influence on the flow distortion.

Combined with the Mach number contours in the symmetry plane (Fig.18), the reasons for the formation of the lowenergy region near the wall of the AIP at θ = 0° can be explained.As can be seen from Fig.18, a low Mach number region appears at the lower part of the bifurcation between the engine duct and the scavenge duct, where the airflow stagnates.Meanwhile, the airflow is divided into an upper stream and a lower stream, which enter the engine duct and the scavenge duct, respectively.The flow entering the main channel flows around the wall from the stagnation point and forms an adverse pressure gradient along the wall.The boundary layer is thus separated under the combined action of the adverse pressure gradient and the airflow viscous force.In this process, vortices are formed in the separated shear layer perpendicular to the streamwise direction (Fig.17), causing a low-energy region close to the wall at θ = 0° at the AIP.

4.3.Effects of attack angle on internal flow structures

Fig.18 Mach number contours and streamline distribution in symmetry plane.

Fig.20 Experimental and numerical wall static pressure distributions on inlet at different angles of attack.

duct is more obvious and greater than that at α = 6°, which leads to the lower inlet performance than that at α = 6°(σ = 0.961 and DC60= 0.094 at α = - 2° while σ = 0.966 and DC60= 0.090 at α = 6°).Fig.22 shows the Q-criterion based vortex structure under the same scale of 1 × 10-5s-2at α = - 2° and α = 6°, where the biggest difference lies in the size of vortices near the wall at θ = 0°, and the vortex is larger when α = - 2°.

5.Conclusions

In this paper, the parametric design method of a turboprop aircraft inlet with scavenge duct is established by extracting and controlling the transition law of the critical characteristic parameters.Design variables include rotation angle and line segment length transition rate, centerline shape, variation law of sectional area,etc.The inlet’s performance and internal flow characteristics are examined by wind-tunnel experiment and numerical simulation.

Within the test conditions,the average total pressure recovery coefficients are all above 0.96 and the circumferential total pressure distortions are less than 0.21.The inlet maintains good performance, validating that the parametric design method proposed in this paper is feasible.

Fig.21 Mach number contours of symmetry plane at α= -2°and α = 6°.

Fig.22 Vortex structure displayed under same scale of Q = 1 × 10-5 s-2.

When the air enters the inlet, the flow direction is divided into two streams,which enter the engine duct and the scavenge duct, respectively.The airflow entering the engine duct assumes a flow trend of rolling-up on both sides under the guidance of the profile, and lee-wake vortices are generated on the back of the power output shaft, which includes a dominant vortex responsible for the total pressure distortion index at the AIP.By comparing the streamlines in the symmetry plane and the vortex structures of the engine duct, it is found that with the increase of angle of attack, the separation shear layer gradually decreases near the inlet wall at θ = 0°, which improves the performance of the AIP.As the angle of attack increases from - 2° to 6°, the performance of the inlet improves gradually,with the total pressure recovery coefficient increased by 0.52% and the distortion decreased by 4.3%.

The parametric design of turboprop aircraft inlet with scavenge duct is realized, but the applicable range of its design parameters needs to be further studied, such as how to select the design parameters to be well compatible with the gearbox and the structure of the nacelle,and how the design parameters affect the flow field and the inlet performance.At present, the aerodynamic experiment without propeller is carried out, and then the internal flow of the inlet with propeller interference will be studied.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This study was co-supported by the Civil Airplane Technology Development Program, China (No.MJ-2020-F-10) and the National Science and Technology Major Project, China (No.HT-J2019-V-0004-0095).