Effect Of Inlet Distortions, Co-swirl And Counter-swirl On Single Axial Fan And Contra-rotating Axial Fans

EFFECT OF INLET DISTORTIONS, CO-SWIRL AND COUNTER-SWIRL ON SINGLE AXIAL FAN AND CONTRA-ROTATING AXIAL FANS Bhaskar Roy, Sq. Ldr. Hari Kumar and Amit Batra Aerospace Engineering Department, IIT-Bombay, Mumbai E-mail: [email protected] Abstract

Introduction

Inlet distortions are flow non-uniformities that result in a decrease in performance and lessening of the operating flow range of a compressor. This paper presents an inlet flow distortion study on a single axial fan and a contra-Rotating Twin rotor axial fan. Experiments were carried out for the following configurations: (a) Clean inlet condition, (b) 90o Circumferential Extent Steady Total Pressure Distortion and (c) A combination of 90o Total Pressure Distortion and localized 90o swirl (both coswirl and counter swirl). Total pressure measurements were done at the inlet and exit of the axial fan/contra-stage, both in the distorted and clean regions. The experimental results were analyzed to obtain distortion parameters like Distortion Index and Axial/Contra-Fan performance parameters like the Pressure Rise coefficient, Mean flow coefficient etc. It is observed that the Distortion Index is highest for the case of combination of 90o Total Pressure Distortion and 90o Localized Counter-Swirl. The severity of the combined distortion is more for a contra-rotating stage as compared to the single axial fan. A significant outcome of the study is the fact that in spite of a higher distortion, counter-swirl produces a lower degradation in performance compared to the co-swirl, which has a lower distortion intensity. Also, the performance penalty in the case of the contra-rotating unit is less as compared to the single axial fan.

One of the most important problems faced by the designers of gas turbine engines is the adverse effect of inlet flow distortions on the engine performance and stability. Extensive theoretical and experimental research has been done on multistage axial compressors, and has yielded marginal improvements in offsetting these adverse effects. The focus has now shifted to the concept of contrarotation, (rotation of two rotors in the opposite direction without any stator), with a view to achieve high pressure rise per stage as well as an improvement on the stability of the stage through rotating stall/surge suppression. The contra-rotation concept has found acceptability in recent developments of fuel-efficient aeroengines. Since inlet distortion is an unavoidable phenomenon, its effect on the design of contrarotating compressors cannot be disregarded. The term “inlet distortion” is used to denote the nonuniformity of any of the flow properties such as total pressure, static pressure, temperature, velocity or flow angle existing at the engine inlet. These have been broadly classified as:

Nomenclature Flow Variables Ca -DC(θ) -Pt -R -r/R -SC(θ) -Um -Ut -α -α -ψ -φ -Post Scripts 1 -2 --

Axial Velocity Distortion Index Total Pressure Blade Tip Radius Radius Ratio Swirl Co-efficient Mean Blade Velocity Blade Tip Velocity Inlet Vane Swirl Angle Flow Angle Pressure rise Co-efficient Mean Flow Co-efficient rotor inlet rotor exit

♦ ♦ ♦ ♦ ♦ ♦

Total pressure distortion, which may be steady or dynamic Temperature distortions Inlet swirl distortions Inlet duct geometry (asymmetric geometry) Back pressure distortions Rotating distortions (geometric and flow related)

Inlet flow non-uniformities arise due to various reasons like changes in aircraft attitudes due to maneuvers, flow separation in air intake due to shock wave/boundary layer interaction, wakes of the aircraft or other aircrafts, vortices, cross winds and atmospheric turbulence, ingestion of hot gases in VSTOL aircraft etc. Regions of localized swirl can be generated due to upstream flow field redistribution. Other causes may be secondary flow effects in the end-wall, or due to inlet vortex problems. The effect of swirl and inlet total pressure distortion appear to interact in a non-linear manner such that their combined effect is more severe than that might be inferred from the simple addition of the effects of swirl and distortion.

Objective In the present study, the effect of inlet distortion flow on the performance of axial Fan and contrarotating twin-rotor axial fan unit was investigated. The experiments were carried out for the following configurations: a) Clean inlet condition b) 90o Circumferential Extent Steady Total Pressure Distortion c) A combination of 90o Total Pressure Distortion and 90o Localized swirl The experiments for the above configurations were carried out on the contra-fan test rig with and without the second (contra) rotor. The localized swirl is produced by deflecting the inlet vane mechanism both at positive and negative incidences. The positive deflection of the inlet vanes results in a circumferential flow in the direction of rotation of the rotor, which is termed as co-swirl in this context. Similarly, the negative deflection produces a circumferential flow that is counter to the rotation of the rotor (the first rotor, in case of the contrarotating fans) and is termed as counter-swirl. The flow measurements were taken using standard Pitotstatic tubes along with the digital micro-manometer. Past Work The effect of inlet distortion on axial compressor performance was recognized in as early as 1950’s, when NACA undertook an extensive experimental investigation of the problem. At the same time, several analytical examinations of the nonaxisymmetric inflow based on the actuator disk model of a compressor blade also appeared1. Since then a large amount of theoretical and experimental work has focused on axial compressors and fans because of their use in military and commercial aircrafts. These efforts have yielded in an overall improvement of the performance of axial compressors but to a limited extent. It is with this background that the interest in contra-rotation concept has emerged with a view to achieve more pressure rise per unit axial length, besides its

suitability (in increasing the operating range) for rotating stall/surge suppression2. A simple form of circumferential non-uniformity is the square-wave distribution, in which two distinct but equal circumferential extents of high and low total pressure regions exist at the compressor face as shown in Figure 1. A basic understanding of the circumferential variation of inlet flow was given by Pearson and Mackenzie3 using the “parallel compressor model”. In this model the compressor is viewed as two identical compressors operating in parallel with different inlet total pressure to a common static pressure. This idea, along with the assumption that each of the two compressors (in parallel) will operate as per the uniform flow compressor characteristic at the local mass flows, allows description of the compressor behavior in the distorted flow. Quantification of Distortion, Swirl and Performance In order to quantify the distortion, the concept of θcrit and distortion index, DC(θ) is used. θcrit is that angle of circumferential distortion which gives maximum loss in surge delivery static pressure. The general trends of compressor performance with different inlet distortions are illustrated in a series of experiments undertaken by Reid4, which are shown in Figure 2a and Figure 2b. Figure 2a shows the compressor delivery pressure at the surge line, for different types of distortion. Figure 2b shows the effect of sub-dividing the total angular extent of the distortion (which is fixed) into different numbers of equal sections. It can be seen that the greatest effect on the loss of surge delivery pressure is observed when there is only one region. Also, as the angular width of the spoiled sector (low inlet total pressure) is increased there is a width above which there is a little change in the exit static pressure. This width is often referred to as the critical sector angle (θcrit). Using this, the distortion index is defined as: DC( θcrit ) = ( Pt |3600 − Pt |worstθcrit )/(1/2)ρ ca

2

DC(θ) can be used to judge both the quality of intake flow and the tolerance of an engine. The distortion index has been found successful in correlating inlet distortion index with compressor performance. The loss of surge margin has been found to be approximately proportional to the distortion index. One of the simplest ways of defining surge margin has been defined by Cumpsty5. For a quantitative measure of swirl, a swirl coefficient SC(θ) analogous to the Distortion coefficient DC(θ) has been defined by Guo6. It is the maximum average circumferential component of the cross-flow velocity in a sector of the measuring station, non-dimensionalized by dividing by the mean duct velocity at the throat section. Figure 1: Square wave non-uniformity

Figure 2: (a) Effect of Circumferential Distortion Angle, (b) Effect of number of sectors on surge pressure ratio4 Experimental and Analytical Studies The effect of inlet flow distortion on compressors can be broken down into two major areas: • The attenuation of the distortion pattern as it passes through the compressor. • The resulting change in compressor performance. One of the early studies related to attenuation was done by Ehrich7 using the actuator disk approach. It was concluded that an attenuation of 0.7 to 0.8 in velocity distortion is attainable and can be improved with lower reaction staging. In another study by Yocum8, it was shown that the attenuation of distortion as it passes through the rotor is a function of the blade stagger angle and the ratio of rotor blade spacing to the distortion wave length. Plourde9 concluded from his studies in a multistage compressor that the overall attenuation of both total pressure distortion and axial velocity distortion is dependent on the slope of compressor pressure flow rate characteristics. The attenuation increases when the slope is made more negative. Stenning10 has done an analysis by the compressor model to predict the effect of circumferential distortion on the performance map. This method has proved useful for analyzing distortions greater than 60o as unsteady effects are neglected in the simple parallel compressor model. An analytical and experimental investigation of asymmetric annulus swirling flows in turbo-machines annuli has been done by Greitzer11. It is found that in a swirling flow the different type of flow disturbances (pressure and vorticity) are strongly coupled. The magnitude of the flow angle and static pressure distortion increases with increasing mean swirl angle and/or decreasing hub-tip ratio. Another study done by Viswanath12 reports on the combined effect of inlet flow distortion and swirl on an axial fan stage. It was found that at the design flow coefficient, swirl causes a deterioration in performance in addition to that caused by distortion. In addition, the attenuation of distortion was high in the presence of swirl. In a recent work, Sharma13 studied the effect of inlet distortion on the performance of the contra-rotating stage using a 360o inlet distortion screen. Other benefits of contra-rotating fan have also been investigated by Roy14,15,16 in past years.

Experimental Set-up The experimental facility consists of the contrarotating fan unit, distortion mechanism, four traverse mechanisms and the instrumentation for measuring the various flow parameters. The Mach number at inlet during these tests was of the order of 0.12, so that effects of compressibility were negligible. The upstream turbulence introduced by the distribution mechanism is approximately 8%. Contra-Fan Test Unit A Schematic layout of the contra-fan test rig used for the present investigation is shown in Figure 3. The first fan comprises of 11 blades followed by ten-bladed contra-fan. The design data of the contrafan stage is given below. The contra-rotating rotors are mounted on two separate solid co-axial shafts. The two rotor shafts independently driven by two separate DC motors are capable of rotating in opposite direction in a speed range of 0-2400 rpm. Mass flow rate can be varied by moving the throttle cone at the rear. Provision is made for the measurement of the flow field at the inlet and at the exit of the contra-fan. The detailed design of the test rig and the contra-fan unit have been provided earlier by the author14. Distortion mechanism Pressure Distortion Generator (Distortion Screen) The basic screen design method involves laying of screens of different porosity over a low blockage base screen. The experimental set up used for the distortion study, consists of an isolated rotor and a twin-rotor contra-rotating fan unit. In the present study only a single screen is used. It is fabricated using a wire mesh of 90o circumferential extent and having a porosity of 0.7. Two MS strips of 2 mm thickness and width 2 cm are cut and bent in a quarter arcs of radii equal to that of hub and outer casing respectively. The wire mesh is cut to a circumferential extent of 90o and soldered between the strips. The distortion screen can be push-fitted between the hub and the casing inner walls. The distortion mesh can be rotated around the annulus and if required, finer meshes can be attached to this base screen to charge the distortion intensity.

Figure 3: Contra Fan Test Rig

1.

2. 3. 4. 5. 6.

7. 8.

9. 10.

Metal strip with slots (for movement of lever) Nut for fixing the lever Pivot pin (outer casing) Inlet vanes Lever to move the inlet vanes Strip fixed on hub for pivoting inlet vanes Outer casing Split pin Pivot pin (hub) Pointer

Figure 4: Isometric View of the Swirl Mechanism Swirl Generator (Inlet Vanes) The inlet vane mechanism to produce localized swirl is shown in Figure 4. It is located just downstream of the distortion screen. The mechanism consists of a series of five plates of 10 cm X 10 cm size arranged in a 90o quadrant of the fan annulus. The inlet vanes are equally spaced in a 90o quadrant and are pivoted at the hub and casing wall. They are rotated by levers which have a slotted pin (5 mm length) soldered on one end of the lever, while the other end is moved in slots cut in a semicircular strip held on the outside casing. The inlet vanes are held in the slotted pin (inserted from outside casing wall) with split pins. This arrangement is used for easy installation and removal of the inlet vane mechanism. The inlet vanes can be rotated from an angle of +15o to –15o. Measuring Stations The details of the measuring stations are given in Figure 5(a) and 5(b). The flow angle measurements

by the three hole probe was done at the mean radius at an axial location of 295 mm from the center of the inlet swirl vanes. The total pressure measurements were taken at two axial locations i.e. inlet and exit of the contra-fan stage and five radial locations corresponding to each location. The measurements were taken in the clean and distorted regions. In the distorted region downstream of the distortion mechanism, the pressure measurements were done using the total pressure rake. The traversing mechanism allowed probe rotation and radial movement. The angle of the total pressure rake is set equal to the flow angle as measured by the three-hole probe to avoid directional errors. Online averaging of pressure is done for the total pressure rake. In the clean flow region, shielded total pressure probes have been used. Two shielded pressure probes were used at the exit of the contrafan stage in both the clean and distorted regions. These were positioned at an axial distance of 185 mm from the second rotor.

1. 2. 3. 4.

5. 6.

7.

Pitot’s static tube Distortion screen Inlet vane swirl Pitot’s rake/ shielded total pressure probe First rotor Contra-rotor Shielded total pressure probe

Figure 5: Measuring Stations (a) axial; (b) radial

Experimental Study, Results and Analysis The experimental data was analyzed to obtain the various fan performance parameters like the pressure rise coefficient ψ, mean flow coefficient φ, Distortion Indices, DC(θ) etc. The different aspects of inlet distortion and its effects on the axial fan performance are summarized below. Inlet Distortion Studies on Single Axial Fan Distortion Indices DC(θ) The distortion indices for the combined distortion for both positive and negative swirl are given in Table 1 at two flow coefficients. It can be seen from the table that distortion index for the case of coswirl is much less than that for the counter-swirl. Since the Distortion Index is a measure of the severity of the distortion, it seems that counter-swirl results in more distortion. Loss of Peak Pressure Rise Coefficient Figure 6 shows the characteristic plot of the axial fan in terms of the pressure rise coefficient, ψ versus the mean flow coefficient, φ for different

configurations. The absolute values of ψ and the percentage reduction in ψ compared to the uniform inlet conditions are given in Table 2. The observations are summarized below: • There is deterioration in peak pressure rise coefficient as the distortion configuration is changed from total pressure distortion to the combined distortion of total pressure and swirl. • It seems that although DC(θ) is higher for counter-swirl, the surge pressure rise achieved is higher compared to the co-swirl which has a lower distortion index. This suggests that there is better attenuation of combined distortion with counter swirl in the axial fan. • It can be seen from the characteristic plot that the curves move downwards and to the left, as the distortion configuration is changed from total pressure distortion to the combined distortion with counter-swirl and with co-swirl, in that order. It is therefore evident that for any meaningful pressure rise under distortion condition, the machine has to operate with lower mass flow rates.

Table 1: Distortion Coefficients for the Combined Distortion

Table 2: Pressure Rise Coefficient (ψ) for various Distortion Configurations

Figure 6: Axial fan characteristics under clean and various distortion conditions Inlet Distortion Studies on Contra-Fan The experimental investigations on the two contrarotating rotors were carried out with a view to assess its performance under the same distortion configurations. The experiments were conducted at three speeds of 2400, 2200 and 2000 rpm. The speed ratio between the two rotors at all the above speeds was kept 1.0. Distortion Indices DC(υ) The distortion indices for the various configurations at the first rotor inlet are given in Table 3. Comparing the values of DC(θ), it can be seen the maximum severity of distortion is that caused by a combination of distortion and counter-rotating swirl. Variation of flow angle, α with inlet vane swirl angle,α, The variation of flow angle, α at mean radius with change of inlet-vane swirl angle α is shown in

Table 4. Two aspects that can be identified are as follows: •

•

The influence of rotor blade suction on incoming flow angle can be seen from Table 4. It can be seen that when counter-swirl is imparted, there is a decrease in the incidence of flow angles at the rotor face. When co-swirl is imparted there is a slight increase in the flow angle at the rotor face. This suggests that the flow angles are more affected by the rotation of rotor in case of counter-swirl as compared to that in case of co- swirl. The second aspect relates to the throttling effect of both the co- swirl and the counter-swirl. Table 4 shows the velocity change at various incidences of inlet vane angle. It seems that the effect of co-rotating swirl is to decrease the mass flow rate. However, in case of counterswirl, this effect of vane blockage is absent.

Table 3: Distortion Indices for various configurations

Table 4: Variation of flow angle with change in inlet vane swirl angle

Table 5: Pressure Rise Coefficient ψ, values at stall onset points

Table 6: Percentage Loss of ψ for different inlet configurations

Figure 7(a), (b) and (c) Contra-Rotating Fan Unit Performance at various speeds and various distortions

Loss of Peak Pressure Rise Coefficient Figures 7(a), 7(b) and 7(c) show the contra-fan stage performance characteristics in terms of the pressure

rise coefficient ψ, and mean flow coefficient φ. The change in pressure rise at instability is expressed as a fraction of the uniform flow pressure rise for the contra-fan stage operating at constant speed. The

stage pressure rise coefficient values at stall onset points for different speed ranges are given in Table 5. From Table 6, it can be seen that, as in the case of single fan, in contra-fan the maximum deterioration in pressure rise coefficient,ψ is in the case of inlet distortion with the co-swirl. This is a very significant observation. The experimental results suggests that in spite of a higher distortion index, DC(υ) in the case of the counter-swirl, the performance degradation is lower as compared to the co-swirl which has a lower distortion index. Comparative Analysis Distortion Index DC (θ) The Distortion Indices for various inlet distortion configurations indicate an increase from total pressure distortion to combined total pressure and swirl. The severity of the combined distortion is much more for counter-swirl as compared to coswirl, both for the axial fan and contra-rotating unit. The deterioration in performance can be observed from the characteristic plots of the axial fan (Figure 6) and of the contra-rotating units (Figures 7a, 7b and 7c). It can be seen that the curves shift downwards and to the left as the distortion configuration is changed from the uniform flow, total pressure distortion and combined distortion, in that order. The maximum deterioration is for the case of distortion with co-swirl. This shift in the characteristic curves implies the following two points: • Under distortion conditions owing to the lower mass flow rates for stable operation of the axial machine, there is a penalty in terms of the pressure rise achieved. The percentage deterioration in pressure rise coefficient is much larger for contra-rotating unit compared to the axial fan. • The distortion attenuation seems to be better for the case of total pressure distortion with counter swirl, as compared to the co-swirl. So, in this study we see that the Distortion Index, which represents the severity of distortion, is higher for a total pressure distortion with counter-swirl compared to co-swirl. One would normally expect a corresponding reduction in the performance. However it is now observed that the pressure rise across the rotors behind the distorted zone is significantly higher in the case of counter swirl as compared to the co-swirl. Conclusion This study presents the effect of 90o Circumferential Extent Steady Total Pressure Distortion and a combination of 90o Total Pressure Distortion and swirl (both co-rotating and counter-rotating swirl), on the performance of an axial fan and a set of twinrotor contra-rotating fans. It is observed from the

experimental results that Distortion Index is highest for the case of combination of 90o Total Pressure Distortion and 90o Localized Counter-Swirl. It is observed that in spite of the higher distortion; counter-swirl produces a lower degradation in performance (in terms of peak pressure rise) compared to the co-swirl. Similar observations have been reported by Fottner17 and Wadia18. The performance penalty (as measured in percentage loss of ψ) for contra-rotating unit is less compared to that of a single axial fan for the co-swirl case. Distortions with counter-swirl produce nearly similar loss of ψ (in %) in both, single and contrarotating fans. In both, single fan and contra fan, distortion with counter-swirl gives lower loss of performance compared to distortion with co-swirl. References 1.

Katz, R., "Performance of Axial Compressors with Asymmetric Inlet Flows", Air Force Office of Scientific Research, USA, TR-59-8, June 1958. 2. Sharma, P.B. and Adekoya, A., “A Review of Recent research in Contra-Rotating Axial Flow Compressor Stage”, ASME Paper, 96-GT-254. 3. Pearson, H. and Mckenzie, A., "Wakes in Axial Compressors", Journal of Royal Aeronautical Society 63, July 1959, p. 415-416. 4. Reid, C., "The Response of Axial Flow Compressors to Intake flow distortion”, 1969, ASME paper 69-GR-29. 5. Cumpsty, N.A., Compressor Aerodynamics, Longman, UK, 1989 6. Guo, R.W. and Seddon, J., "Swirl Charactersitics of S-Shaped Air Intake with Both Longitudinal and Vertical Offsets", Aeronautical Quarterly, May 1983. 7. Ehrich, F., "Circumferential Inlet Distortions in Axial flow turbomachinery", Journal of the Aeronautical Sciences, AIAA, June 1957, pp. 413-417. 8. Yocum, A.M. and Henderson, R.J. "Effects of Some Design Parameters of an Isolated Rotor on Inlet Flow distortions", Journal of Engineering for Power, ASME Transactions, Vol. 102, 1980, pp. 178-186. 9. Plourde, G.A., "Attenuation of Circumferential Inlet distortion in Multistage Axial Compressors", AIAA Journal of Aircraft, Vol. 5, No. 3, 1968,. 10. Stenning, A.H., "Inlet Distortion Effects in Axial Compressors", Journal of Fluids Engineering, ASME Trans., Vol. 102, 1980, pp. 7-11. 11. Greitzer, E.M. and Strand, T., "Asymmetric Swirling flows in Turbomachine Annulii", Journal of Engineering for Power, Vol. 100, 1978, pp. 618-629.

12. Vishwanath, K. and Govardhan, M., "Effect of Circumferential Inlet Distortion and Swirl on the Flow Fluid of an Axial Flow Fan Stage", ASME, Vol. 8, 1996, pp. 96-GT-253. 13. Sharma, P.B., “Development of ContraRotating Axial Flow Compressor”, research report no. IITD/ME, 86001, IIT-Delhi, New Delhi. 14. Roy, B.; Ravibabu, K.; Rao, S.P.; Basu, S.; Raju, A. and Murthy, P.N., “Flow Studies in Ducted Twin-Rotor Contra-Rotating Axial Flow Fans”, ASME 92-GT-390, 1992. 15. Roy, B. and Agrawal, L., “Casing Boundary Layer Control by Recess Vaned Casing for a Twin-Rotor Contra-Rotating Axial Flow Fan Unit”, ASME 94-GT-478, 1994. 16. Kumar, S. and Roy, B., “Endwall Flow Development Across a Contra-Rotating Fan Unit”, ASME 2000-GT-0502, 2000. 17. Fottner, L.; Jahnen W.; Peters Thomas. "Effects of 90o Extent Circumferential Inlet Distortion on Aerodynamic Performance of a Axial Compressor Stage", ASME 99-GT-440 , ASME Turbo-Expo, 1999, USA. 18. Wadia, A.R., Wolf, D.P.; Haaser, F.G.; “ Aerodynamic Design and Testing of an Axial Flow Compressor with Pressure Ratio of 23.3:1 for the LM2500+ Gas Turbine”. ASME 99GT-210, ASME TURBO-EXPO, 1999, USA.