Paper ID: ETC2019-110 Proceedings of 12th European Conference on Turbomachinery Fluid dynamics & Thermodynamics ETC13, April 8-12, 2019; Lausanne, Switzerland

OPEN ACCESS Downloaded from www.euroturbo.eu

1 Copyright © by the Authors

IMPROVEMENT OF THE PERFORMANCE OF AN AXIAL FAN WITH COUNTER-ROTATION

A. Ghenaiet 1 – I. Beldjilali 2

1 Laboratory of Energetics and Conversion Systems, Faculty of Mechanical Engineering, University of Sciences and Technology Houari Boumediene, BP32 El-Alia Bab-Ezzouar, 16111, Algiers, Algeria,

Email: ag1964@yahoo.com 2 Laboratory of Turbomachinery, Fluid Mechanics and Energetic, Ecole Militaire Polytechnique,

BP17 Bordj-el-Bahri, 16046, Algiers, Algeria, Email:beldjilali.brahim@gmail.com

ABSTRACT The counter-rotation solution is preferred to increase the power and flow rates of fans.

The aims of the present work are to experimentally and numerically characterize the aerodynamic performance of a counter-rotating fan (CRF) and investigate the effects of some design parameters such as the variable stagger angle and inter-distance. The baseline axial fan was modified by adding a rear rotor, a spherical nose and a tail fairing, thus a higher pressure rise is obtained from the conversion of a large part of swirl flow into a static pressure. The results reveal the effects of stagger angle and axial inter-distance on the improved aerodynamic performance and show the best setting parameters. Besides, the counter-rotation seems reducing the aerodynamic loading of both rotors and allowing working at a stable static pressure rise over a wide range of air flow rate.

KEYWORDS

Counter-rotating axial fan, CFD, Aerodynamic performance, Stagger angle, Inter-distance

NOMENCLATURE Roman letters c chord [m] CP pressure coefficient d inter-distance [m] N rotational speed [rpm] P power [W] P pressure [Pa] P pressure rise [Pa] Qv flow rate [m3/s] Re Reynolds number T torque [m.N] V forward speed [m/s]

Greek letters efficiency angular speed [rps]

dynamic viscosity [kg/m.s] density [kg/m3] ξ stagger angle [deg]

Subscripts s shaft t t-s t-t

static net power total total-to-static total-to-total

Acronyms CRF FR RR SRF

counter-rotating fan front rotor rear rotor single-rotating fan

INTRODUCTION The adoption of the solution of counter-rotation has opened a way towards designing high

performance turbomachines for various applications. The main reasons are the improvement of pressure rise and efficiency by recovering the kinetic energy from the front rotor (FR), besides of improving flow capacity in both ducted and unducted configurations (Toge and Pradeep 2015). At

inlet of reaconvertingstator stage2013). AnSharma et affected byand axial sgives bette(Roy et al. revealed a stall. ShigeCRF, and efficiency characterisreveal the pressure riverified exinvestigatedesign conof FR. Misof a high aand the axpresent stuinvestigateand inter-d

AXIALThe ba

blades of mfrom C4 pr156.5 mm,configuratito recover of RR is gexcept the triangles an

A test-

flow-meterthrottling. arrows indpressure at

ar rotor (RRg it to a state, RR not o

n early expal. (1988)

y several facspacing. Preer off-desig1992). Punsignificant emitsu et ashowed thais higher a

stic becomeeffect of v

ise. The loxperimentalled numericanditions, andstry and Praaspect-ratio xial spacingudy is to she the influendistance betw

L FAN CONaseline singmodified starofile. The , respectiveions of SRFthe tangent

generated bastagger an

nd the radia

F

rig (Fig. 2)r, FR and RThe two rotdicate the t inlet is de

R) the tangtic pressure

only recovererimental i

). They havctors such aeliminary stgn performandhir and Shimproveme

al. (2010) cat the mainat partial fle accentuatevarying rotaower rotatioly by Furukally as well d found highadeep (2013and low-sp

g have a sighow the ponce of certaween FR an

NFIGURAle-rotor axiagger angle,blade stagg

ely and the F and CRF tial energy ased on the gles were d

al equilibriu

(aFigure 1: A

was built tRR connecte

tors and elelocations o

educed from

ential compe rise leads rs the static investigatiove concludeas the speedtudies by diance and thharma (1992ent of off-decompared thn branch offlow rates ced. Nouri eational speeonal speed kawa et al. (

as experimher efficien3) carried opeed CRF. Tgnificant rossibility of ain parametnd RR.

ATIONS ANial fan rotat, having a m

ger angle is tip clearancwith incorpby RR, thersame profi

defined accoum design.

a) Axial fan co

to test both ed via an intectric motorof measuremm the avera

2

ponent contto saving ihead but a

n on a 0.6ed that the d ratio of FRfferent rese

he possibilit2) tested a design perforhe charactef CRF characompared tet al. (2012ed of both and its ad

(2000, 2007mentally thency and operout a paramThey have vole on the f designing ters on the

ND TESTS tes at a fixemaximum chmodified. Tce is equal porated sphreby the floiles of FR. ording to th

nfiguration

SRF and Ctermediate drs are suppoments for ge of static

tributes to ain consumelso supplies

66 hub-tip performan

R and RR, bearchers havty of increaducted counrmance, parristics betwacteristic iso SRF, but) developedrotors over

dvantage on7) and Noure performanrating marg

metric study verified thataerodynamia low-speeperforman

ed speed ofhord lengthThe tip and to 2.2 mmerical nose

ow is suppoThat means

he flow dat

ns: a) SRF

CRF, whichduct (in bluorted by strustatic press

c pressure ta

a higher reled power. Cs energy to ratio compce of a cou

blade staggeve shown thasing the opnter-rotatingrticularly in ween simples stable witt at low flod a low sper a wide rann the perforri et al. (201nce of CRF gin for the h

on the aerot the speed ic performa

ed CRF fromce, specific

f 1440 rpmh of 117.5 m

hub diamet. Figure 1 sand tail fai

osed to exit s all the dima of FR, co

(b) and b) CR

h is composeue) and a lonuts to the exsure and teapings at p

lative velocCompared to

the fluid (Wpressor is runter-rotatinering, pitch-hat the counperating rang compresso

the contexte axial fan th a negativow rates theed CRF tenge of flowrmance of 12). Chen eunder desig

highest rotatodynamic pratio of the

ance. The am a baselin

cally the sta

m. This rotormm which ar

ters are 479shows the tiring. CRF axially. Th

mensions aronsidering t

RF

ed of a Venng dischargxternal casinemperature.oint (P1) an

city, and byo the rotor-Wang et al.,reported byng stage is-chord ratio

nter-rotationnge as wellor stage andt of rotating(SRF) and

ve slop andhe slope ofest setup tow rates and

CRF wereet al. (2008)gn and off-tional speederformance

e two rotorsaims of thene fan, andagger angle

r has sevenre designed9.6 mm andtwo studiedis designed

he geometryre the samethe velocity

nturi nozzlege duct withng. The red. The totalnd the flow

y -,

y s o n l d g d d f o d e ) -d e s e d e

n d d d d y e y

e h d l

w

3

rate, which is quite reasonable since the air velocity is low. On the other hand, the outlet total pressure from SRF or CRF is obtained from the static pressure tapings far downstream at point (P4), to allow for diffusing the circumferential flow momentum, and by adding the dynamic component and the ducting losses between (P3) and (P4) the effective outlet total pressure is determined. The airflow was measured by means of a Venturi nozzle manufactured according to BS848 (1997) which was calibrated for this purpose. The net shaft power for the direct drive is deduced from the consumed electrical power of the three-phase electrical motors reduced losses (BS4999, 1987). The measurements of voltage, current, frequency, speed of rotation and slip of induction were made. Figure 2 does not show the instrumentation. The measurement uncertainties depended on instrumentation for which the absolute errors were: 2.45 Pa for static pressure difference, 0.2 K for temperature, 0.1 rpm for rotational speed, and about 3.5 W for electrical power.

Legend P1 : Location for static pressure and temperature at the throat of venture-meter. Also the location of Pitot tube P2 : Location for static pressure upstream the front fan Pi : Location for static pressure half a distance between FR and RR P3 : Location for static pressure and temperature downstream of FR P4 : Location for static pressure allowing recovering the flow rotation component in addition to the temperature

Figure 2: Test-rig

The measured performance of SRF for two stagger angles are shown by (Fig. 3), depicting that the pressure rise decreases with the flow rate, and more at low staggering. In fact the flow rate is restricted by the test-rig ducting and throttling, but the real flow range is larger at low stagger angles. The maximum air speed at the venturi-meter throat is about 15.7 m/s. Figure 4 presents the absorbed and net shaft power to drive the fan, which are compared for different stagger angles. It appears that the consummed electrical power increases at low stagger angles but has a slight variation with the flow rate, on the contray the net shaft power depends strongly upon the flow rate. The lesser shaft power corresponds to the highest stagger angle of 82 deg which reveal a peak value, while for the other stagger angles no peak is seen since the flow range was not totaly swept because the throttling imposed. The cconfiguration of stagger angle ξSR = 62 deg seems offering the highest efficiency, which will be further discussed.

(a) (b)

Figure 3: Measured total and static pressure rise of SRF: a) ξSR=72°; b) ξSR=62°

Venturi-meter

RR FR

Inter-space

Throttling

4

(a) (b)

Figure 4: Measured power of SRF: a) Absorbed; b) Net shaft power

The measured pressure rise of CRF (Fig. 5) depicts improved performance and broader flow range compared with SRF. RR has a positive effect as a lesser power is consumed compared with the tandem fans. This is obvious by comparing between the operating point at the upper limit of flow rate 1.05 m3/s and a stagger angle of 72 deg, showing that the power consumed by SRF is 210 W (Fig. 4(b)) while that of CRF is about 400 W (Fig. 6). For the optimal stagger of 62 deg this benefit is much evident, where the power consumed by SRF is 275 W and that of CRF is 375 W, which is a noticeable gain. By referring to the nominal flow rate, for the configuration of CRF (ξFR=62°-ξRR=67°) the gain in efficiency reaches a value of 0.82 instead of 0.75, while the static and total rises to 275 Pa and 355 Pa, respectively, instead of 125 Pa and 185 Pa for SRF, and in parallel the flow range is broader. By changing inter-distance in between 160 - 360 mm, which is in tem of chord distance equal to 1.5c, 2.5c and 3.5c, FR influences the power consumed by RR due to flow interaction between them. For the configuration of CRF (ξFR=62°, ξRR=67°) and inter-distance d=1.5c the total-to-static efficiency is around 64.94% and the shaft power is at the maximum and the static pressure rise is equal to 324 Pa. On the other hand, for the larger inter-distance d=3.5c the static pressure rise drops moderately to 301 Pa, since the shaft power of RR is less owing to the mixing having an effect on tangential velocity and dynamic pressure between the two rotors and subsequently the pressure loading of RR. As consequence the total-to-static isentropic efficiency varied slightly about the value of 64.7%.

(a) (b)

Figure 5: Measured static pressure rise of CRF for different inter-distance (d=1.5c, 2.5c, 3.5c) and stagger angle: a) ξFR=72°- ξRR=76°; b) ξFR=62°-ξRR=67°

Figure 6: a) ξFR=72°

NUMEThe Re

are solved the high re10-6 for conthe mixingsubsonic tutotal tempeoutlet. Theguide vanethe lateral angles of (ξFR=52°-ξtypical com

The m

treatment avolumes dik-SST tuReynolds m<y+<100).

parametermean chor

Consumed°- ξRR=76°;

ERICAL STeynolds Avby the cod

esolution adnvergence. g plane (sturbomachinerature at t

e inlet free se span. All t

sides. FourFR and RRξRR=59°). Mmputational

mesh generaat the blade istortion an

urbulence mmodel in fi

The first

, ∞

rd. The dim

(a) d power of

b) ξFR=62°

TUDY veraged Navde Ansys-CFdvection schThe frozen

tage interfanery, the robthe inlet whstream turbuthe walls arr configuraR, successi

Moreover, thdomain is r

Figure

ation adoptetip to captu

nd resolvingmodel the aune mesh renode from

2⁄ , and frmensionless

CRF for i

°-ξRR=67°

vier-Stokes FX. The k-heme is useinterface is

ace) is usedbust boundahich correspulence intenre of non-slations of CRively: (ξFR=he inter-distrepresented

e 7: Compu

ed non-unifure flow detg the boundautomatic ne

egions (y+< m a wall ∆

friction factos paramter

5

inter-distan

(RANS) eqSST turb

ed with a los used to invd to compuary conditiopond to me

nsity is selip conditionRF are stud=82°-ξRR=8tance d is v

d by Fig. 7.

utional dom

form distribtails. Arounary layers. Fear-wall tre2) and the

∆

is

or 0.03 is less

nce (d=1.5c

quations of bulence modocal time scvestigate thute the aerons consist easured valet at 5 % ann and adiabdied, which4°), (ξFR=7

varied to 1c

main for th

bution of hnd the bladeFigure 8 pr

eatment allowall-functiplaced ac

7 . functhan 85. T

(b) c, 2.5c, 3.5c

the flows tdel (Menter

cale factor ohe flow detarodynamic in applyinglues and a nd the lengthbatic. The peh are define72°-ξRR=76°c, 1.5c, 2c a

he CRF

hexahedral es O-grids alesents the g

ows switchiion in regio

ccording to

ction of The mesh s

c) and stag

through SRr1994) is adof 10 and a ails. On the performanc

g the total pmass flow h scale to 0eriodicity is

ed by differ°), (ξFR=62and 3c (c: c

blocks withllowed mingrid of FR bing betweenons of high

the non-d

∞c⁄ bize depend

gger angle:

RF and CRFdopted, andresidual ofother hand,

ce. In suchressure andrate at the

.1 times thes applied torent stagger°-ξRR=67°),chord). The

h a specialnimizing theblade. Withn the low-values (20

dimensional

ased on thedency study

F d f , h d e e o r , e

l e h -0 l

e y

6

conducted at 1440 rpm and Qv=0.6 m3/s, considering the isentropic efficiency (Fig. 9), permitted adopting a total number of 473282 nodes for SRF and 846062 nodes for CRF.

Figure 8: Rotor meshing Figure 9: Grid dependency

RESULTS AND DISCUSSIONS The computation of aerodynamic performance is based on the stage interface, where the

circumferential mass averaging of static and total pressure and temperature are done at the inlet and outlet, whereas the torque is integrated over the blade surfaces. The total-to-static and total-to-total isentropic efficiency is given for: SRF by:

∆ Q

T and

∆ Qv

T (1) and (2)

and CRF by:

∆ Q

TFR TRR and

∆ Q

TFR TRR (3) and (4)

Validations The first validation is done for SRF, but only the stagger angle ξSR=62° is shown by Fig. 10.

Owing to the inaccuracies of instrumentation and the limited flow range due to throttling imposed by the test-rig there is an acceptable agreement between the computed and measured performance. However, some important discrepancies are observed in both the pressure rise and efficiency in the middle of flow range. In particular, at the most, a difference of about 8 points in efficiency is visible due to higher losses induced by the front and rear struts and the build-up of boundary layer at the casing and also attributed to the power mismatch. The second validation concerns CRF, which for the configuration (ξFR=62°-ξRR=67°) shows (Fig.11) a fairly good consistency with the measurement, and the performance and flow range are improved compared to SRF.

(a) (b)

Figure 10: Compared performance of SRF (ξSR=62°): a) Static pressure rise; b) Total-to-static isentropic efficiency

7

(a) (b) Figure 11: Compared performance of CRF (ξFR=62°-ξRR=67°): a) Static pressure rise; b) Total-to-static isentropic efficiency

Computed Performance of SRF Figure 12 present the computed static pressure rise and total-to-static isentropic efficiency for

SRF at different stagger angles and flow rates, revealing strong dependence upon the stagger angle. The first branch exhibit a rapid variation and a negative slope followed by another branch of a moderate variation ending by curve bend of negative slope. This latter for the lowest stagger angle of 52 deg is steepest leading rapidly to the unstable region. At stagger ξSR=62° the characteristics reveal higher pressure rise and a peak in total-to-static efficiency equal to 64.7 %, in addition the flow range goes up to 2.1 m3/s beyond the limit reached experimentally.

(a) (b) Figure 12: Computed performance of SRF: a) Static pressure rise; b) Total-to-static isentropic efficiency

Computed Performance of CRF Figures 13 present the computed performance of CRF for different stagger angles and axial

inter-distances, depicting gains in the pressure rise and isentropic efficiency compared to SRF. Over the major part of operating range the slope of pressure rise is negative which means a good working stability. It seems that the lowest stagger setting (ξFR=52°-ξRR=59°) is characterized by lesser stable operating range. The CRF configuration (ξFR=62°-ξRR=67°) is shown to produce the best performance, where the static and total pressure rise are about 229 Pa and 336 Pa, which are more than that obtained for SRF. Besides, the total-to-static efficiency was initially equal to 64.7% for SRF is improved to 71.2 % for CRF, as explained by the significant increase in static pressure rise and lesser power consumed. Moreover, CRF operates at a wider range of high efficiency compared to SRF since the RR provides a better suction.

8

(a)

(b)

(c)

(d)

Figure 13: Static pressure rise and total-to-static isentropic efficiency of CRF compared with SRF: a) ξFR=82°-ξRR=84°; b) ξFR=72°-ξRR=76°; c) ξFR=62°-ξRR=67°;d) ξFR=52°-ξRR=59°

9

Table 1 summarizes the measured and predicted performance for the different configurations of SRF and CRF. The obtained performance is given in terms of total and static pressure rise and isentropic efficiency, which correspond to each nominal operating point defined at the maximum isentropic efficiency. Table 1 does not show a comparison between experimental and numerical values at the same flow rate. Indeed, there is a difference in the flow range between the measurement and the computation added to the respective optimum isentropic efficiency which are shifted, and this why the relative error is not provided.

Table1: Comparison between performance at the nominal operating points of SRF and CRF

Parameters

FR Stagger angle82 deg 72 deg 62 deg 52 deg

SRF CRF SRF CRF SRF CRF SRF CRF Qv (m

3/s) -Num 0.61 0.64 1.18 1.22 1.45 1.87 1.47 2.51 -Exp 0.60 0.51 1.03 0.92 1.45 1.55 1.21 1.73 ΔPS (Pa) -Num 112 233 100 248 66 229 74 229 -Exp 117 269 120 294 112 324 131 294 ΔPt (Pa) -Num 135 248 153 293 172 336 171 424 -Exp 148 340 159 341 190 354 235 404 t-s (%) -Num 53 59.3 63 69.8 64.7 71.2 36.5 69.4 -Exp 46.57 52.6 51 53.91 51.5 64.94 30.5 64.5 t-t (%) -Num 64.2 60.5 73 72.5 78.33 77.7 51.3 78.8 -Exp 58.59 59.7 67.57 64.3 74 74.95 54.71 71.5

Flow Structures The discussion of the flow field corresponds to the nominal point of CRF (ξFR=62°-ξRR=67°)

and inter-distance d=1.5c. As shown by Fig.14, the flow streamlines through the two rotors follow more or less the rotor blades and reach the highest velocity over the suction side of RR. At the entrance of second rotor there is an increase in relative velocity from root to tip with a variation in flow direction. At exit from RR the flow is practically axial at the upper part of blade, but the flow rotation exists in the lower part. Figure 15 depicts the static pressure which continues increasing through RR. This latter has a dual contribution since it continues the diffusion started in FR and adds more energy to the fluid. Figure 16 plots the pressure coefficient ( and are the

local and static pressure at entry is the relative inlet velocity) around the two rotors blades, when CRF operates at the nominal condition. The blade staggering has a significant effect on aerodynamic loading due to higher suction of blades. For FR and RR the pressure loading is more significant over the upper part of blade, and FR is more loaded. By comparing between CRF and SRF the pressure loading is evidently lesser for FR, which is good for the stability since the pressure gradient over the suction side is alleviated.

Figure 14: Streamlines coloured by flow velocity of CRF (ξFR=62°-ξRR=67°, d=1.5c)

10

Figure 15: Static pressure of CRF (ξFR=62°-ξRR=67°, d=1.5c)

(a) (b) (c)

Figure 16: Pressure coefficients for (ξFR=62°-ξRR=67°, d=1.5c°: a) FR, b) RR; c) SRF (ξSR=62°)

Axial Velocity Profiles Downstream of FR a swirl velocity is established to ensure a radial equilibrium balancing the

axial velocity. From hub to mid-span axial velocity tends to reduce the flow incidence, but in contrary incidence increases from mid-span to tip due to blockage. Figure 17 shows the profiles of axial velocity halfway between FR and RR, for both high and low stagger angles. Since the flow at exit from FR is assumed to follow the free vortex design method, the tangential velocity multiplied by the radius is constant along the radial direction. For both configurations the averaged axial velocity seems having the expected form between 20-80% of span, in addition the flow is discharged mainly through the upper span. The defect in axial velocity seen at outermost part is due to the casing boundary layer and tip clearance effects. At the lower part, the negative value of axial velocity indicates a recirculation. It may be concluded that the presence of RR affects the axial velocity profile by reducing it near hub, while increasing it practically on the remaining of blade span. This latter indicates a contraction effect, which can be explained by the aspiration effect of RR which interacts with the shroud boundary layer and the hub vortex.

(a) (b)

Figure 17: Axial velocity between FR and RR: a) ξFR=72°-ξRR=76°; c) ξFR=62°-ξRR=67°

11

The design of RR assumed that the inlet velocity is the same as that at outlet of FR and ignored the mixing and decay. Figure 18 presents the velocity profiles downstream of RR for both high and low inter-distances. Near the tip region the axial velocity is relatively lower in the inter-space than downstream of RR, which could be attributed to interaction of FR tip vortex and RR, which was also observed by Cho et al. (2009). Besides, the axial velocity downstream of RR and near the hub is negative owing to flow separation and hub vortex. When the fluid passes through RR, the axial velocity increases somehow below 30 % of span and continues increasing in between 30 - 80% of span, and decreases above 80 %.

(a) (b)

Figure 18: Axial velocity downstream RR: a) ξFR=72°-ξRR=76°; c) ξFR=62°-ξRR=67°

Influence of Inter-Distance The flow entering RR is assumed identical to that from FR, but the axial inter-distance has a

significant influence on the flow structure to a certain limit. According to Zachcial and Nürnberger (2003) and Hetherington and Moritz (1977) this distance allows the mixing to occur prior to entering the next rotor passage. Additionally, the tip flow from FR has a significant impact on the casing boundary layer of RR. The previous figures show that the velocity profiles become non-sensitive to inter-distance beyond d=1.5c, but below .i.e. d=1c, the influence is more evident. It seems that beyond a certain distance the axial velocity profile is not influenced except near the hub which could be attributed to flow recirculation. The flow angle between RR and FR is influenced by inter-distance over a wide portion of blade height. Below 20 % of blade height, the deflection angle of FR increases when inter-distance increases from 1.5c to 3.5c, and the important deflection at trailing edge causes extra losses. The flow angle relative to the blade of RR decreases too much in the tip and root regions, and thus the flow deflection and aerodynamic loading of FR reduce substantially. For the configuration (ξFR=82°-ξRR=84°) the distance d=1c offers a total-static efficiency about 59.3%, and by lowering the stagger angle this distance would be farther due to flow interactions. For the optimal configuration (ξFR=62°-ξRR=67°) an axial distance d=1.5c seems producing the highest total-to-static efficiency which is about 71.2 % at the air flow rate of 1.87 m3/s. The same distance applies for the configurations (ξFR=72°-ξRR=76°) and (ξFR=52°-ξRR=59°). Besides, there is an optimal distance which beyond the net shaft power supplied to RR will increase and subsequently the efficiency drops.

CONCLUSION The aerodynamic characterization of SRF and CRF allowed conclusions about some design

parameters, mainly the blade staggering and inter-distance, while keeping the same speed of rotation. There are noticeable improvements, which besides increasing the pressure rise and the isentropic efficiency to a certain extent, the flow range is also improved due to better suction imposed by RR. The configuration (ξFR=62°-ξRR=67°) with an inter-distance around d=1.5c, gives the best performance. Furthermore, the aerodynamic loading is lesser in CRF compared to SRF, and the pressure gradient over the blade suction side is alleviated for FR, which is good for the stability

12

reason. Further understanding how FR and RR interact at various axial gaps is an important step to determine the beneficial effects of counter-rotation and confirm the design choices.

REFERENCES BS4999 PART3, (1987). Testing of Electrical Motors, 1987 BS848, (1997). Fans for General Purposes. Part 1. Performance Testing Using Standard Airways.

Amendment 1997 Chen, Y.Y., Liu, B., Xuan, Y., Xiang, X. R., (2008). A study of speed ratio affecting the performance of

a counter-rotating axial compressor. Proc IMechE, Part G: J Aerospace Engineering 2008; 222: 985–991. Cho, L. S., Cha, B. J., Cho, J. S., (2009). Experimental study on the three-dimensional unsteady flow

characteristics of the counter-rotating axial flow fan. Journal of Fluid Science and Technology, Vol.4, No.1:200–209, 2009.

Furukawa, A., Cao, Y., Okuma, K., Watanabe, S., (2000). Experimental study of pump characteristics of counter-rotating axial flow pump. Proc. 2nd Int. Symposium on Fluid Machinery and Fluid Engineering, Beijing, 67-657, pp. 245-252.

Furukawa, A., Shigemitsu, T., Watanabe, S., (2007). Performance test and flow measurement of counter-rotating axial flow pump. Journal of Thermal Science, Vol. 16, No. 1, pp. 7–13.

Hetherington, R., Moritz, R. R., (1977). The influence of unsteady flow phenomena on the design and operation of aero engines. Unsteady Flow Phenomena inTurbomachinery. AGARD CP. 177, North Atlantic Treaty Organization

Mistry, C., Pradeep, A. M., (2013). Effect of variation in axial spacing and rotor speed combinations on the performance of a high aspect ratio counter-rotating axial fan stage. Proceedings of IMechE, Part A: Journal of Power and Energy 2013 227: 138.

Nouri, H., Ravelet, F., Bakir, F., Sarraf, C., Rey, R., (2012). Design and experimental validation of a ducted counter-rotating axial-flow fans system. ASME Journal of Fluids Engineering, Vol. 134, No.10, 104504, doi:10.1115/1.4007591.

Patankar, S. V., Spalding, D. B. A. (1972). Calculation procedure for heat, mass and momentum transfer in three-dimensional parabolic flows. Int. J. of Heat and Mass Transfer, vol.15, pp. 1778-1806, 1972.

Pundhir, D. S., Sharma P. B., (1992). A Study of aerodynamic performance of a counter-rotating axial compressor stage. Defence Science Journal, Vol 42, No 3, July 1992, pp 191-199

Roy, B., Ravibabu, K. Srinivasa Rao, P. Basu, S. Raju, A., Murthy, P. N., (1992). Flow studies in ducted twin-rotor counter rotating axial flow fans. Paper No. 92-GT-390, pp. 1–8, Proceedings of the International Gas Turbine and Aero Engine Congress and Exposition, June 1–4, 1992, Cologne, Germany,

Sharma, P. B., lain, Y. P., lha, N.K., Khanna, B. B., (1985). Stalling behavior of a counter-rotating axial compressor stage. Paper No. ISABE 85-7087. In Proceedings of the 7th ISABE, 2-6 September 1985, Beijing.

Sharma, P. B., Jain, Y.P., Pundhir, D. S., (1988). A study of some factors affecting the performance of a counter-rotating axial compressor stage. Proceedings of IMechE, Part A: Journal of Power and Energy 1988 202: 15

Shigemitsu, T., Fukutomi, J., Okabe, Y., (2010). Performance and flow condition of small-sized axial fan and adoption of counter-rotating rotors. Journal of Thermal Science, 19, 1–6, Springer

Sullivan, T.J., (1990). Aerodynamic performance of a scale model, counter-rotating unducted fan. ASME Journal of Turbomachinery, Oct.,Vo1. 112, pp.579.

Toge, T. D., Pradeep, A. M., (2015). Experimental investigation of stall inception mechanisms of low speed counter rotating axial flow fan stage. International Journal of Rotating Machinery, Volume 2015, Article ID 641601, 14 pages

Wang, J., Ravelet F., Bakir, F., (2013). Experimental comparison between a counter-rotating axial flow fan and a conventional rotor-stator stage. 10th European Turbomachinery Conference, 15-19 April 2013, Lappeenranta, Finland.

Zachcial, A., Nürnberger, D. (2003). A numerical study on the influence of vane-blade spacing on a compressor stage at sub- and transonic operating conditions. Paper GT2003-38020, pp. 807-818; 12 pages, ASME Turbo Expo 2003, Atlanta, Georgia, USA, June 16–19, 2003