SYSTEM-LEVEL ASSESSMENT OF TAIL-MOUNTEDPROPELLERS FOR REGIONAL AIRCRAFT

Roelof Vos∗ and Maurice F. M. Hoogreef∗∗Delft University of Technology, Faculty of Aerospace Engineering, Delft, The Netherlands

Keywords: Aircraft Design, Propulsion Integration, Ducted Propellers, Technology Assessment

Abstract

Three regional transport aircraft of different con-figuration are synthesized for the same designspecification using an automated design routine.The first aircraft features wing-mounted pro-pellers, the second aircraft features propellersmounted on the horizontal tail plane, while thelast configuration replaces the horizontal and ver-tical tail with two ducted propellers mounted nearthe rear of the fuselage. These last two innova-tive configurations have the potential to reducethe cabin noise, while the ducted propeller couldalso reduce community noise. The analysis anddesign methods to size and analyze these con-figurations include weight and balance, stabilityand control, aerodynamic performance, and mis-sion performance. Propeller slipstream effectsare taken into account and demonstrated to playan important role in the sizing of the horizon-tal tail surface. A comparison study between thethree aircraft for a harmonic mission of 1530kmand 7500kg payload demonstrates that the air-craft with wing-mounted propellers has the low-est maximum take-off mass and burns the leastamount of fuel. The two innovative configura-tions have slightly less performance, which is ul-timately attributed to the large center-of-gravityexcursion that stems from an aft-mounted propul-sion system. A 3% increase in maximum take-off weight is predicted along with a fuel burnincrease between 5% and 10% for the innova-tive configurations, respectively. Further inves-tigation of the underlying assumptions might im-prove these results in future studies.

1 Introduction

This paper presents the conceptual design andassessment of two innovative aircraft config-urations that rely on tail-mounted propellersand fuselage-mounted, ducted propellers, respec-tively. By positioning the propellers further be-hind the main cabin of the aircraft, the cabinnoise stemming from the propeller and the tur-boshaft engine can be substantially reduced in-creasing the level of comfort of the passen-gers. This configuration was already proposedin the 1980s [3] and is currently being investi-gated in the Clean Sky 2 project called IRON[5]. While the tail-mounted propellers reducethe cabin noise, the flyover and sideline noisehardly change. As a matter of fact, care mustbe taken when installing the propeller to preventadditional noise stemming from the wing vortexsheet entering the propeller disk at elevated an-gles of attack and low flap settings. To reducethe community noise, the second innovative con-figuration shields the propeller with a cowling.This cowling (sometimes termed “duct” or “ringwing”) replaces the vertical and horizontal tailplanes and therefore has a synergistic effect: itshould reduce the community noise, it should im-prove the propulsive efficiency, and it should pro-vide sufficient stability and control authority overthe entire center-of-gravity excursion. To demon-strate that the configuration with a ducted pro-peller is feasible a flight test on a subscale modelwas performed in 2016 (see Fig. 1).

While the flight test was successful and theaircraft demonstrated excellent handling qualities

1

ROELOF VOS AND MAURICE F. M. HOOGREEF

Fig. 1 Subscale test article of aircraft withfuselage-mounted ducted propellers prior todemonstration flight.

during powered flight, it left the question openwhether such an innovative configuration wouldalso improve the fuel burn of the aircraft com-pared to a more conventional, high-wing aircraftwith its propellers mounted on the wing. There-fore, this paper presents a system level assess-ment of three regional aircraft that are each de-signed for the same mission requirements buthave a different propulsion-airframe configura-tion. The main objective is to find out whatthe difference is between these aircraft configura-tions is in terms of their weight and balance, theiraerodynamic performance, and their mission per-formance.

The following sections will describe themethodology that is followed to design these air-craft (Sec. 2), the top-level aircraft requirementsand assumptions (Sec. 3), and a discussion of theresults (Sec. 4).

2 Methodology

The conceptual designs of the aircraft are madeusing the Aircraft Design Initiator (termed theInitiator). The process flow of the Initiator isshown schematically in Figure 2. On the di-agonal, the various modules of the Initiator areshown (grey boxes), while the white boxes indi-cate the variables that serve as output and/or in-put to the various modules. The large blue ar-rows indicate the feed forward and feedback di-rections, while the three colored square perime-

ters mark the three (partially nested) convergenceloops that are present in the Initiator. It should benoted that the variables that are presented hereoften represent an umbrella for a set of variables.For example, "Fuselage geometry" includes vari-ables on the fuselage cross section, length, up-sweep angle, etc. Furthermore, it should also benoted that the synthesis process is a process ofconvergence, where the design variables are al-tered in an iterative way until a predefined setof performance indicators converge below a cer-tain threshold within a large set of constraints. Inother words, the Initiator uses a process of design”feasilization” [8], rather than optimization to geta converged aircraft design.

Figure 2 only shows the process flow on anaggregated level. Many of the blocks on the diag-onal line contain multiple design, analysis or siz-ing modules. For example, "Geometry Modules"contains more than 20 individual modules that di-mension the aircraft geometry, ranging from en-gine position to wing taper ratio. Other modulesare not listed in Figure 2 such as the sizing mod-ules to determine the horizontal tail or vertical tailarea. More information on the process flow of theInitiator can be found in Ref. [2]

For the design of aircraft with (ducted)tail-mounted propellers, particular modificationswere made to the stability and control module aswell as to the weight and balance module. Theseare described in more detail in the subsection be-low. The propeller slipstream effects on lift andpitching moment were included according to themethod of Obert, described in Ref. [1].

2.1 Weight Estimation Ducted Propellers

As the weight estimation of the propulsion sys-tem plays a pivotal role in the assessment of thevarious calculations, an overview of these equa-tions is presented here. All masses are computedin pounds (lbs). The engine is integrated in thenacelle, which is connected to the fuselage bymeans of a pylon. The total nacelle mass, includ-ing the pylon, consists of three components: thecenterbody (cb) around the turboshaft engine, the

2

SYSTEM-LEVEL ASSESSMENT OF TAIL-MOUNTED PROPELLERS FOR REGIONAL AIRCRAFT

Fig. 2 Design Structure Matrix of the Initiator process flow. On the diagonal the main modules areshown, while on the off-diagonal cells the input/output of modules is displayed. The colored boxesindicate the three nested loops that the initiator executes iteratively.

duct around the propeller, and the pylon:

mnac = mcb +mduct +mpylon (1)

The centerbody around the turboshaft engine iscomputed according to:

mcb = 0.14WTO (2)

The duct mass is computed according to the na-celle formula for turbofans provided by Raymer[6] assuming 60% of the nacelle mass can be at-tributed to the duct:

mduct = 0.40l0.1ductw

0.294duct n0.119

lim m0.611ps N0.984

en Sduct

Nen(3)

where lduct and wduct are the length and width ofthe duct (both in ft), respectively, nlim is the ul-timate load factor, Nen is the number of engines,and Snac is the duct outer surface area in ft2. Theducted propellers are connected to the fuselageby means of pylons. To estimate their mass, theestimation from Torenbeek [9] for horizontal tail

planes has been used assuming they are fixed andthat they have no sweep:

mpylon = Spylon

(3.81S0.2

pylonVD−0.287)

(4)

where, VD is the dive speed in kts, Spylon is theplanform area of the pylon in ft2.

2.2 Aerodynamics of Ducted Propellers

Aerodynamic analyses for the ducted propellerin power-on and power-off conditions were de-rived from wind tunnel tests performed in thelow-turbulence tunnel of TU Delft and reportedin the MSc thesis of Harinarain.[4] The referencearea that was chosen for the coefficients reportedin this subsection equals the projected area of thetwo ducts onto a horizontal plane:

Sduct, ref =2D

ARduct(5)

where D is the duct diameter, and ARduct is theaspect ratio of the duct (D/c), which is a designchoice.

3

ROELOF VOS AND MAURICE F. M. HOOGREEF

The maximum lift coefficient, CLh of the ductwhen the propellers were not turning was foundin the wind tunnel tests to be to 1. When deflect-ing the internal control surfaces, an increase inmaximum power-off maximum lift coefficient of20% was found. Furthermore, at a propeller ad-vance ratio of 0.28 and 0.42 the maximum liftcoefficient was measured to be 5.1 and 2.5, re-spectively without any control surface deflection.Based on these results, this study assumes an un-powered maximum lift coefficient of CLh = 1.2and a (conservative) CLh = 3 for maximum powerconditions.

For the angle-of-attack derivative, CLα, first

the power-off lift-curve slope is computed basedon the ring-wing (rw) model of Weissinger [10]:

CLα|rw =

π

2ζclα (6)

where clα is the lift-curve slope of the airfoil usedin the ring wing and ζ is computed as follows:

ζ =1

a+λπ

2 +λ tan−1 (1.2λ)(7)

where λ = 1/ARduct. The lift-curve slope of theducted propeller system increases when thrust isapplied according to the following formula:

CLα=CLα

|rw (1+ kT Tc) (8)

where Tc is the thrust coefficient of the propellerwith respect to the projected area of a single ductand kT is multiplication factor that was found tobe 0.2 [4].

3 Design Specifications

To assess the effect of tail-mounted propulsionon the aircraft performance, the aircraft are de-signed according to the top-level aircraft require-ments (TLARs) that are stated in Table 1. Theserequirements are similar to the specifications ofan ATR-72-600, which serves as a reference air-craft in this study.

Apart from the TLARs, the Initiator usesmany requirements on performance, stability andcontrol, and geometric clearances that stem from

Table 1 TLARs for regional aircraft design.Spec. Unit ValueRange km 1530Structural payload kg 7500Pax - 68Cruise altitude m 7000Cruise Mach - 0.45Time to climb to 5400m min. 17.5

the certification specifications (CS-25), the mil-itary specifications (MIL-F-8785), or handbooks[7]. The harmonic mission profile, for which theaircraft is sized, consists of the following phases:engine start and warm-up, taxi, take-off at sealevel, climb and accelerate to cruise altitude andcruise speed, cruise at constant Mach number andaltitude, decent and decelerate to sea level andapproach speed, climb to 3300m and travel 100km (diversion), decent to 1300m, loiter for 30minutes at 1300m, decent to sea level, land andtaxi out. To be able to fly over a longer rangewith less payload it is assumed that a fuel tank ispresent between the front and rear spars of eachwing half, spanning 80% of the semi-span, start-ing at the wing symmetry plane.

Each aircraft is sized with a constant aspectratio of 12. Each passenger is assumed to have amass of 80kg and to carry 22kg of luggage. Theremaining payload mass (564kg) is assumed tobe additional cargo. All luggage and cargo areassumed to be stowed below the passenger cabinas bulk. The passengers are seated in a 2-2 con-figuration with a single aisle and a 26 inch seatpitch. Each aircraft is assumed to be equippedwith high-lift devices that can increase the (un-powered) maximum lift coefficient to CLmax =3.2. Furthermore, a thrust-specific fuel consump-tion based on the cruise thrust is assumed of14.5g/kN/s and a constant propeller efficiency of80% is assumed for cruise, 65% for take-off and75% for climb. Furthermore, the induced thruststemming from the lip suction of the engine cowl-ing (if present) is neglected in the computations.

4

SYSTEM-LEVEL ASSESSMENT OF TAIL-MOUNTED PROPELLERS FOR REGIONAL AIRCRAFT

4 Results

Using the TLARs from Section 3 and themethodology of Section 2, three aircraft weresynthesized using the Initiator: 1) a high-wing,T-tailed aircraft with wing-mounted propellers(WMP), 2) a low-wing aircraft with propellersmounted on the tips of the horizontal tail plane(TMP), and 3) a low-wing aircraft with fuselage-mounted ducted propellers (FMDP). The three-views of these synthesized aircraft designs areshown to-scale in Figure 3. From these figuresit can be observed that the fuselage of each air-craft has the same outer mold line (OML), whilethe wings of the two aircraft with aft-mountedpropulsions systems have a slightly larger wing,which is shifted further aft on the fuselage. Alsothe undercarriage of these aircraft is further aftand attached to the wing, making the need for afuselage fairing obsolete. From the front viewit can also be seen that the number of propellerblades varies between the three different configu-rations, although this had only a very minor effecton the overall results.

4.1 Wing Loading and Power Loading

As the Initiator sizes the aircraft by selecting adesign point with the highest wing loading andat that wing loading the highest power loading,it is useful to inspect the wing-loading versuspower-loading diagram displayed in Figure 4.Only the constraints that actively bound the de-sign space are shown, omitting the constraints forsecond-segment climb gradient, time-to-climb,and buffet-onset, which are not active. It can beseen that the take-off field length, the one-engine-inoperative climb gradient requirement in land-ing configuration, and the landing distance arethe requirements that size each of these aircraft.It can also be seen that the cruise constraint isquite different for all three configurations. Thisis due to the difference in the untrimmed dragpolars between the three configurations as willbe elaborated below. The wing loading of theWMP, TMP, and FMDP is 3330, 3340, and 3350N/m2, respectfully, while the power loading is

66.7, 64.6, and 63.0 N/kW, respectively. Thismeans that although the design point for each air-craft is virtually the same for each of the aircraftin terms of power loading and wing loading, theamount of cruise excess power is highest for theWMP and lowest for the FMDP configuration.

4.2 Weight and Balance

The weight distribution of the three empty air-craft is schematically shown in Figure 5. Thelocation of heavy components such as pressurebulkheads, undercarriage, APU and engines canclearly be discerned. Note that the engine loca-tion is visibly different between the three differ-ent aircraft configurations, placing a large mass atthe tail of the aircraft for the TMP and the FMDP.The coinciding circles at the engine location areindicative of the engine mass, the cowling massand the engine system mass. While the dry en-gine mass for both the TMP and the FMDP areboth 600kg, the cowling (or duct) of the FDMPis estimated to add 200kg to the installed engineweight. This is still less than the estimated 470kgof the vertical and horizontal tail plane weight ofthe WMP aircraft. It should be stressed, that theweight of the duct is merely an estimate as no ref-erence data is available on the structural weightof ring wings or propeller ducts.

As the Initiator positions the wing such thatthe lowest horizontal tail plane area results, eachaircraft has a different wing position. The re-sulting CG excursions for each of the aircraftwith respect to their respective mean aerody-namic chords are shown as a function of the load-ing in Figure 6. These loading diagrams showthree lobes for each aircraft. The lowest twolobes can be attributed to the loading of the pas-sengers in window seats and aisle seats. The up-per lobe is for the luggage and cargo. It can beseen that the WMP aircraft is assumed to have asingle cargo bay below the main passenger floorand therefore its loading diagram looks differentfrom the low-winged configurations that are as-sumed to have a forward and aft cargo bay, sepa-rated by the center wing box.The two ascendinglines represent the CG shift resulting from fuel

5

ROELOF VOS AND MAURICE F. M. HOOGREEF

0 5 10 15 20

scale (m):

WMP TMP FMDP

Fig. 3 Three-views of the three synthesized aircraft designs: a conventional aircraft with wing-mountedpropellers (WMP, left), a low-wing aircraft with tail-mounted propellers (TMP, center), and a low-wingaircraft with fuselage-mounted ducted propellers (FMDP, right).

2000 2400 2800 3200 3600 4000

Take-off Wing Loading, W/S [N/m2]

0

0.05

0.10

0.15

Tak

e-off

Pow

er L

oad

ing, W

/P [

N/W

]

sL=

915 m

sTO = 1333 m

Mcr= 0.45 (c/V)= 2.1 %

Design Point

Reference Aircraft

bm

ax=

36 m

TMP

FMDP

WMP

Fig. 4 Wing-loading versus power-loading dia-gram for each of the three synthesized aircraftconfigurations.

burn in the case of the harmonic mission (left)and the ferry mission (right). Also the bounds ofthe CG diagram are indicated as vertical lines in-tersecting with the horizontal axis of the diagram.

From Figure 6 it can be observed that the

loading diagram of the WMP aircraft is quite fa-vorable. The total center of gravity (CG) excur-sion measures 30% of the MAC, while for theTMP and the FMDP it measures 40% and 41%,respectively. Furthermore, the CG diagram of theWMP aircraft is much less skewed, which keepsthe CG around the 25% meaning that little trimdrag is to be expected from this configuration.The TMP and FMDP aircraft have their quarter-chord points of the MAC at 49.2% and 52.2% ofthe fuselage length, respectively. This results ina fairly skewed CG shift with for both aircraft.The reason for this is the large propulsion sys-tem mass that is located relatively far behind thewing. To balance and stabilize the aircraft, thewing therefore shifts further aft and the operat-ing empty mass (OEM) CG is found at approx-imately 54% and 47% of their respective meanaerodynamic chords (MACs). During loading ofpassengers the CG consequently shifts forwardresulting in a total CG excursion of 0.41c and0.40c for the TMP and FMDF, respectively.

6

SYSTEM-LEVEL ASSESSMENT OF TAIL-MOUNTED PROPELLERS FOR REGIONAL AIRCRAFT

-1

5

20015-5

100 5

0

2

4

6

10

0

2

4

6

10

5200

15-5 10-10 5

0

TMP FMDPWMP

0

2

10

5

20015

-510

-10 5

Fig. 5 Distribution of mass over the aircraft. Each circle represents a point mass where the area of thecircle is representative of the mass of the individual components.

Relative CG location, x/c [~]

12

14

16

18

20

22

24

Air

craf

t m

ass,

m [

t]

0 0.1 0.2 0.3 0.4 0.5 0.6

TMP

FMDP

WMP

Fig. 6 Loading diagrams of the three differentaircraft configurations.

4.3 Stability and Control

Figure 7 shows all the constraints that relatethe tail area ratio (Sh/S) to requirements on bal-ance and stability. If we look at the WMP air-craft we can see that the sizing requirementsare the power-off equilibrium in landing condi-tion at most forward CG, i.e. pitch-up to CLmax ,and the power-on stability in take-off configura-tion at most aft CG. The effect of the propellerand its slipstream causes a destabilizing tail-offpitching moment derivative. As the T-tail is notin the slipstream of the propeller, its effective-ness is not influenced by the power of the pro-peller. This results in a slightly larger tail arearatio of 28% compared to the scenario where pro-peller slipstream effects are neglected. It can also

be seen that, while not sizing, the take-off rota-tion constraint and the power-on pitch up con-straints are also close to the feasible design space.This is because the slipstream effect in power-onconditions generates a relatively large nose-downpitching about the main gear wheel axle and for-ward center-of gravity, respectively.

If we examine the scissor plot of the TMP(lower left in Fig. 7), it can be seen that thepower-on constraint lines on the equilibrium sideof the plot are all absent. Even though, theseconstraints were evaluated, they fall outside therange of the curve to the increased effectivenessof the horizontal tail plane in power-on condi-tions. Rotation during take-off, or pitch-up toCLmax could therefore be satisfied with a muchsmaller tail, provided the power would alwaysbe on. As this is not the case, the tail surfaceis merely sized for power-off conditions. Withthe required CG shift of 41%, this results in tailarea ratio of 37%, which are lower than the 42%that was found my Goldsmith in his study intothis configuration as alternative to the DC-9 [3]but larger than the 35% assumed by Nicolosi etal. for a 130-seat aircraft with the same configu-ration [5].

The stability plot of the FMDP shows thatthe horizontal tail area ratio should be 31%. Thehorizontal tail area for this aircraft is defined asthe sum of the projected areas of the two ductsonto the XY plane. The required tail size there-fore effectively sizes the diameter of the duct, as-suming a duct aspect ratio (D/c) of 2. It canbe seen that the tail is sized for power-off con-

7

ROELOF VOS AND MAURICE F. M. HOOGREEF

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

Relative CG location, x/c [~]

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

HT

P S

urf

ace

Are

a R

atio

, S

h/S

[~

]

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55H

TP

Su

rfac

e A

rea

Rat

io, S

h/S

[~

]

Relative CG location, x/c [~]

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

WMP

TMP FMDP

Equilibrium landing (power-off)

Equilibrium landing (power-on)

Rotation takeoff (power-on)

Stability cruise (power-off)

Stability cruise (power-on)

Stability takeoff (power-off)

Stability takeoff (power-on)

Stability landing (power-off)

Stability landing (power-on)

CG Excursion

Fig. 7 Scissor plots demonstrating compliance with stability and balance requirements for each aircraftconfiguration. The arrows show the CG excursion for each aircraft configuration.

ditions, which is to be expected as the tail ef-fectiveness increases with increasing propulsivepower. This increases both the maximum lift co-efficient of the ducts and their lift-curve slope.When no power is present, the ducts act as ringwings which are consequently sized to meet allconstraints on balance and stability. The result-ing dimensions of the duct (including structuralthickness) are an outer diameter of 5.0m and achord length of 3.3m. The propeller that is fit-ted in the ring wing has a resulting diameter of4.6m. A redesign of this aircraft might favor aduct with a smaller aspect ratio to reduce the pro-peller diameter and improve the noise shieldingcapability of the duct.

4.4 Aerodynamic Performance

The aerodynamic performance of the three dif-ferent aircraft configurations is shown in Figure8. For each polar, a straight line has been addedthat originates at the origin and is tangent to thedrag curve. The tangency point shows the lift co-efficient for which the highest aerodynamic effi-ciency can be obtained. It can be seen that allthree drag polars have approximately the sameshape, which is to be expected as they all have thesame aspect ratio, wing incidence and fuselage.However, it can also be seen that the WMP hasthe lowest minimum drag value of 211 counts,while the TMP and FMDP have 219cts, and235cts, respectively. Note also that the wing area

8

SYSTEM-LEVEL ASSESSMENT OF TAIL-MOUNTED PROPELLERS FOR REGIONAL AIRCRAFT

(and therefore reference area) of each of the air-craft is different and measures 68.1m2, 70.7m2,and 70.5m2 for the WMP, TMP, and FMDP, re-spectively. For the TMP and the FMDP, the in-crease in minimum drag coefficient, combinedwith the larger reference area therefore resultsin an even higher increase in drag during cruisecompared to the WMP.

Drag coefficient, CD [cts]

0

0.5

1

1.5

Lif

t co

eff

icie

nt,

CL [

~]

100 200 300 400 500 600 700 8000

TMP

FMDP

WMP

(L/D)max

Fig. 8 Drag polars of the three different aircraftconfigurations at M = 0.45.

For the TMP, the increase is attributed to thelarger HTP and VTP that result from the aft-mounted engine. The total wetted area of theHTP plus VTP increases from 66m2 (WMP) to86m2 (TMP). For the FMDP, the large increasein minimum drag coefficient is also attributed tothe friction drag of the propeller ducts. Because itis assumed that both the outside and inside of thering wings contribute to the friction drag coeffi-cient, the total wetted area of the two ring wingsadds up to 205m2, which is more than double thewetted area of the TMP and triple the wetted areaof the WMP. In practice, it is anticipated that partof this friction drag is offset by an increased lipsuction over the ring wing that is present whenthe propellers provide thrust as shown by Hari-narain [4]. However, this has not been consideredin the present study.

Finally, one can conclude from the drag po-lars of Fig. 8 that the optimum cruise lift coeffi-

cient for maximum L/D is around 0.8 for each ofthese aircraft. At the set cruise altitude of 7000m,each of these aircraft cruise at a lift coefficient of0.57, which is below their predicted optimum.

4.5 Mission Performance

The conventional aircraft with wing-mounted en-gines is very similar to the ATR-72-600. In fact,its estimated maximum take-off mass (22870kg)is close to the reference value (22800kg). Also,the operating empty mass is relatively close(13200kg versus 13300kg, respectively). Whilethe wing loading is 11% lower than the refer-ence value (3330 N/m2 versus 3670N/m2, re-spectively), the power loading is 5% higher(0.064N/W versus 0.061N/W, respectively). Thisindicates that the assumed maximum lift coeffi-cient in landing condition (CLmax = 3.2) is slightlyunderestimated. The length of the fuselage mea-sures 24.8 meters, which is considerably smallerthan the 27.1m of the reference aircraft due tothe lower slenderness ratio of the tail cone thatwas chosen, the low seat pitch that was selected,and the fact that luggage and cargo are assumedto be stowed below the passenger floor. Thequarter-chord point of the mean aerodynamicchord (MAC) is located at 46.5% of the fuselagelength, which is a little aft of the reference valueof 43%.

While all three aircraft fulfill the same setof requirements, their key performance indica-tors differ substantially, as is shown in Table 2.The two aircraft with aft-mounted propulsionssystems show an increase in maximum takef-off mass of 1.0 and 0.8 metric tonnes, respec-tively. The TMP aircraft has the highest increasein weight, due to the increased mass of the fuse-lage and horizontal tail plane. However, due toits more favorable CG position at its maximumzero-fuel weight it has less trim drag and there-fore burns less fuel compared to the TMDP air-craft.

The harmonic mission sizes each of these air-craft. The obtained mission profiles are shown inFigure 9 and show overall similarity between thethree different configuration except for the climb

9

ROELOF VOS AND MAURICE F. M. HOOGREEF

Table 2 Key performance indicators for the threesynthesized aircraftParameter Unit WMP TMP FMDPMTOM t 22.9 23.7 23.7OEM t 13.2 13.8 13.7Mission fuel t 2.2 2.3 2.5Power loadingN/kW 63.7 63.5 63.0Wing loading N/m2 3330 3340 3350Xc/4/Lfus - 0.47 0.49 0.52

profile which is dictated by a predefined splitof excess power between acceleration and climb.The time-to-climb requirement of 17.5 minutes isnot reached due to a lack in excess power. It canbe observed that the harmonic mission is com-pleted after approximately 3 hours and 10 min-utes.

Time, t [hrs]

5000

Alt

itude,

h [

m]

and R

ange,

R [

km

]

Altitude

Range

6000

7000

4000

3000

2000

1000

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

TMP

FMDP

WMP

Fig. 9 Resulting mission profiles for each of thethree synthesized aircraft configurations: Wing-mounted propellers (WMP), tail-mounted pro-pellers and fuselage-mounted ducted propellers(FMDP).

While the on-design mission performance isvirtually identical for all aircraft, the payload-range diagram shown in Figure 10 shows thatthe off-design mission performance differs be-tween the three configurations. Due to the su-perior aerodynamic performance, the WMP air-craft has the highest off-design range, regardlessof the payload weight, the TMP comes in second,while the FMDP has the lowest off-design range.It should be noted that each of these aircraft issized with a fuel tank spanning up to 80% of their

relative wing span. As the TMP and FMDP havea 4% larger wing area, their fuel tank is about5% larger then for the WMP. So even as they cancarry more fuel, their off-design mission rangeis still lower than for the WMP. This attributedto two factors: weight and friction drag. Dueto the increased operating empty weight, the in-duced drag is higher at any point in the cruiseflight. Furthermore, the larger minimum drag co-efficient stemming from the increase in wettedarea also increases the parasite drag of the air-craft. As a result, the WMP aircraft has the bestoff-design mission performance.

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

Range, R [km]

12

14

16

18

20

22

24

Air

craf

t m

ass,

m [

t]OEM

Payload

TOM

TMP

FMDP

WMP

Fig. 10 Comparison of the payload-range dia-grams for the three different aircraft. TOM istake-off mass, OEM is operational empty mass.

5 Conclusions

Two innovative aircraft configurations featuringtail-mounted propeller and fuselage-mounted,ducted propellers have been assessed in terms oftheir key performance indicators assuming theyare designed for a harmonic mission of 1530kmand 68 passengers. A conventional aircraft withwing-mounted propellers has been designed as areference case and showed very comparable char-acteristics to the ATR-72-600 in terms of geom-etry and weight. It has been shown that the twoinnovative configurations can be designed to thesame top-level aircraft requirements as the con-ventional aircraft and that these designs couldbe feasible in terms of stability, balance, and

10

SYSTEM-LEVEL ASSESSMENT OF TAIL-MOUNTED PROPELLERS FOR REGIONAL AIRCRAFT

point performance. The innovative aircraft areexpected to see an increase in maximum take-off mass of 3%, an increase in operating emptymass up to 5%, and an increase in fuel burn upto 12% over the harmonic mission. Each of thesechanges has been attributed to the large center-of-gravity excursion during loading that is caused bythe aft location of the propulsion system for theseinnovative configurations. While the aircraft withducted propellers showed the least favorable re-sults, its drag estimates were fairly conservativeand the beneficial interaction between fan andduct were not yet taken into account. This couldoffset some of the drag increases reported hereinand make the ducted propeller a viable alternativeto the wing-mounted propeller configuration.

References

[1] BOUQUET, T., AND VOS, R. Modeling thepropeller slipstream effect on lift and pitchingmoment. In Proceedings of the 2017 AIAAAerospace Sciences Meeting (Grapevine ,TX,2017).

[2] ELMENDORP, R., VOS, R., AND LA ROCCA,G. A conceptual design and analysis method forconventional and unconventional airplanes. InProceedings of the 29th Congress of the Inter-national Council of Aeronautical Sciences (St.Petersburg, Russia, 2014).

[3] GOLDSMITH, I. M. A study to define theresearch and technology requirements for ad-vanced turbo/propfan transport aircraft. Tech.rep., 1981.

[4] HARINARAIN, V. Aerodynamic performancestudy on ducted propeller system for propulsionand control and stability applications. Master’sthesis, 2017.

[5] NICOLOSI, F., CORCIONE, S., TRIFARI, V.,CUSATI, V., RUOCCO, M., AND DELLA VEC-CHIA, P. Performance evaluation and doc es-timation of an innovative turboprop configura-tion. In Proceedings of the 2018 Aviation Tech-nology, Integration, and Operations Conference(Atlanta, GA, 2018).

[6] RAYMER, D. P. Aircraft Design: A ConceptualApproach. AIAA, 2012.

[7] ROSKAM, J. Airplane Design. DAR corpora-tion, 1986.

[8] SCHUT, J., AND VAN TOOREN, M. De-sign "feasilization" using knowledge-based en-gineering techniques. Journal of Aircraft 44(2007), 1776–1786.

[9] TORENBEEK, E. Synthesis of Subsonic Air-plane Design. Delft University Press, 1982.

[10] WEISSINGER, J. Zur aerodynamik desringfluegels. die druckverteilung duenner, fastdrehsymmetrischer fluegel in unterschallstroe-mung. Zeitschriftfiir Flugwissenschaften 4(1956), 141.

Acknowledgements

The authors would like to acknowledge the con-tribution of ir. Tom Schouten, ir. Vikesh Hari-narain, and ir. Malcom Brown whose work wasused to generate the results presented in this pa-per.

Contact Author Email Address

Corresponding author:Roelof Vos; R.Vos@tudelft.nl

Copyright Statement

The authors confirm that they, and/or their companyor organization, hold copyright on all of the origi-nal material included in this paper. The authors alsoconfirm that they have obtained permission, from thecopyright holder of any third party material includedin this paper, to publish it as part of their paper. Theauthors confirm that they give permission, or have ob-tained permission from the copyright holder of thispaper, for the publication and distribution of this pa-per as part of the ICAS proceedings or as individualoff-prints from the proceedings.

11