Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685 www.springerlink.com/content/1738-494x

DOI 10.1007/s12206-013-0416-0

An investigation on the effects of irregular airfoils on the aerodynamic performance of small axial flow fans† Li Zhang1,2, Ying-Zi Jin1,* and Yu-Zhen Jin1 1

The Province Sky Lab of Fluid Transmission Technology, Zhejiang Sci-Tech University, Hangzhou, China Department of Application & Engineering, Zhejiang Economic & Trade Polytechnical, Hangzhou, China

2

(Manuscript Received November 23, 2012; Revised February 27, 2013; Accepted March 4, 2013) ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Abstract In this paper, the effects of airfoils on the aerodynamic performance of small axial flow fans were investigated to reduce airflow turbulence on the blade surface and to improve the aerodynamic performance of small axial flow fans. Irregular airfoils where several convex grooves are bound in the blade pressure surface of the fans have two kinds. The wave-shaped edge is bound to the blade trailing edge of the fans and is designed from the smooth airfoil of the fans. The filtered N-S equations with the finite volume method and the standard kε turbulence model were adopted to carry out the steady simulation calculation. The large eddy simulation and the FH-W noise models were adopted to carry out the unsteady numerical calculation and aerodynamic noise prediction. The results of simulation calculation are in good agreement with the tests, which proves that the numerical calculation method is feasible. The spectrum characteristics of aerodynamic noise of the smooth airfoil and the two kinds of irregular airfoils were analyzed. Although the fans of the three airfoils are regarded as noise sources, the vortex distribution features in the unsteady flow field are also described. Noise reduction mechanisms of the irregular designs of the airfoils were also discussed. The results of this research may provide proof of the parameter optimization and the structural design of small axial flow fans with low noise. Keywords: Large eddy simulation; Aerodynamic noise; Vortex; Noise reduction; Irregular design; Performance ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

1. Introduction A fan is widely adopted in various industrial applications, such as home appliance and personal computers, because of their reliable and simple structures. A dominant trend in the physical design of electronic devices is the continuing rise in device heat dissipation, and this trend is driven by a rapid rise in circuit densities and a push towards much smaller device footprints. One of the drawbacks of this trend is an associated increase in radiated acoustic noise. Conventional noise control methods such as fan speed reduction are effective in reducing fan noise. However, solutions for flow rate reduction are inherently limited [1]. One way to reduce noise may be the use of an active noise control. Studies have shown that active control is an effective method to reduce low-frequency blade noise [2-4]. However, the effective band of active control is very narrow. Thus, the application of active control schemes to solve noise problems is usually fraught with problems. Optimizing fan blade structure or airfoil is considered to be an ideal noise reduction method because aerodynamic noise *

Corresponding author. Tel.: +86 571 86843348, Fax.: +86 571 86843367 E-mail address: [email protected] † This paper was presented at the ISFMFE 2012, Jeju, Korea, October 2012. Recommended by Guest Editor Hyung Hee Cho © KSME & Springer 2013

generated by fan blades is the main noise source of small axial flow fans. For example, many authors designed the leading edge of swept blades by reducing aerodynamic noise and improving the range of operating conditions [5-7]. The bionic blades generated by silencer features of several animal organs produced good results in noise reduction [8-10]. In recent years, scholars have fabricated irregular designs of fan blades. You Bin [11] studied the internal flow characteristics of two axial flow fans with a diameter of 300 mm and used in an outdoor air conditioner unit. Dimples were carved on the suction sides of the blades of one fan. Numerical results showed that the aerodynamic characteristic and static pressure of the fans did not improve when dimples were carved on the suction sides of blades. The dimples on the suction sides of the blades increased the noise in a lower volume flow rate condition, and did not influence the aerodynamic noise at a large volume flow rate condition. A study [12] used a hollow blade from the hub to the trailing edge at the tip of the blades with a diameter of 500 mm to reduce the adverse flow conditions at the hub and the tip of the blades. The results showed a reduction of the adverse flow conditions near the trailing edge at the tip of the blade. Studies [13-16] also used the numerical simulation and experimental methods to investigate the influence of toothshaped blades on aerodynamic noise and internal flow condi-

1678

L. Zhang et al. / Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685

Fig. 1. Sample fan. Fig. 3. Computational domain.

(a) Airfoil B

(b) Airfoil C

Fig. 4. Mesh.

Fig. 2. Irregular designs of airfoil A.

tions. The results showed that the tooth-shaped blades could reduce the aerodynamic noise of the fans and weaken the rotor wake. A study [17] designed a coupling bionic fan with a serrated and strip structure, and the results showed that the aerodynamic noise generated by the coupling bionic blade was lower than that generated by the traditional blade. In this paper, two irregular designs were applied to the smooth airfoil of the fans where several convex grooves are bound to blade pressure surface of the fan and a wave-shaped edge is bound to the blade trailing edge of the fan. The aerodynamic noise and the internal flow feature of the three airfoils were comparatively analyzed, and the irregular designs of the blades on aerodynamic performances were discussed.

2. Description of the fans A small axial flow fan, RF24S9225H, was selected as the sample fan in the paper. Its smooth blades are defined as airfoil A, as shown in Fig. 1. The diameter of the impeller is 84 mm, and its thickness is 18 mm. It has a hub ratio of 0.39, a tip clearance of 1.5 mm, seven blades, a blade stagger angle of 46.9°, a rated rotating speed of 3000 r/m, a rated air volume of 0.010 kg/ s, and a rated air pressure of 37.25 Pa. Two irregular designs were applied to the airfoil A where several convex grooves are bound in blade pressure surface of the fans (airfoil B) and a wave-shaped edge is bound to the blade trailing edge of the fans (airfoil C), as shown in Fig. 2. From Fig. 2(a), the spacing between the convex grooves is equal to each other, and the distance is 8 mm. The distribution of the convex grooves shows a concentric annular arrangement. The diameters of the convex grooves are 4 mm. From Fig. 2(b), the distribution feature of the wave-shaped trailing edge is the denser blade root and the looser blade tip.

Fig. 5. Grid sensitivity test result.

3. Numerical procedure 3.1 Computational domain The computational domain for the simulation consists of the rotating fluid area, the surrounding pipeline, and the inlet and outlet ducts. For simulation purposes, the lengths of the inlet and outlet ducts were extended. The centre of the fan hub was set as the coordinate origin. The length of the outlet extension is 400 mm, and its diameter is 200 mm in the computational domain, as shown in Fig. 3. 3.2 Mesh The Gambit software was used to generate the grid for the computational domain. A non-structural grid (tetrahedral Tgrid) was used in the rotating fluid area and in the surrounding pipeline, whereas a structural grid (hexahedral cooper mesh) was used in the regions of the inlet and outlet ducts, as shown in Fig. 4. The total grid number was over 2.0×106 for each of the fans of the three airfoils, which was justified by the griddependence study. The degree of twist of the grid is predominantly between 0.1 and 0.5. Large grids were used in flow simulation, as shown in Fig. 5, to verify if the solution is grid independent or not. A grid

L. Zhang et al. / Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685

1679

larger than 1200000 does not produce a visible difference in the converged flow rate. 3.3 Boundary conditions The inlet boundary condition is set as a given flow rate. For the outlet, the total pressure is taken as the boundary condition. The solid walls, such as the vane surfaces and the hub, satisfy the no-slip condition in the computational domain. At the rotor-stator interfaces of the fan rotor and the casing, moving mesh approaches were applied to consider the influence of rotor-stator interaction. 3.4 Governing equations

1-Test fan rotor 3- Torque-speed sensor 5-Gauge screen 7- Pressure tap

2- Static pressure tap 4- Driving motor 6-Nozzle 8-Booster fan

Fig. 6. Schematic view of test rig and its components list.

The calculations were performed with a commercial software FLUENT. This code uses the filtered Navier-Stokes equations with the finite volume method. The basic equations that describe the flow include the conservation of mass and momentum. In general form, the conservation of mass in direction xi ( i =1, 2, 3) at time t is given by ∂ρ ∂ + ( ρ ui ) = 0 ∂t ∂xi

(1)

where ρ is the density and ui is the velocity in direction i . The conservation of momentum is described by ∂ ∂ ∂p ∂τ ij ( ρu ) + ( ρu u ) = − + i i j ∂t ∂x ∂x ∂x i i j

(2)

2×10-5 s (equivalent to 1000 time-steps per impeller revolution). The time-step was able to capture the main unsteady phenomena that were observed in the experiment. Noise prediction can be carried out only when the statistically steady solution is acquired in the LES of an unsteady flow field. The FW-H equation based on the Light-hill acoustic analogy theory was used to simulate sound production and dissemination when the noise was predicted [19]. The noise spectrum distribution from the axial fan was obtained after the source data was obtained by integrating the time domain and when fast Fourier transform (FFT) processing was applied. The time-step of the noise calculation is set to 2×10-5 s, and the cutoff frequency of noise calculation is 20 Hz.

4. Experimental measurements where p denotes the pressure and τ ij is the stress tensor. 3.5 Solving methods 3.5.1 Steady calculation The Mach number of the airflow is under 0.3, hence, the fluid is regarded as incompressible. An implicit method of solving the incompressible segregated equation and the RNG k-εturbulence model were used to solve the steady simulation. The SIMPLC algorithms were used to solve the coupling of velocity and pressure. A second-order upwind discretization was used for convection terms, and central difference schemes were used for diffusion terms. 3.5.2 Unsteady and acoustic calculation The large eddy simulation (LES) method was used to predict the completely unsteady flow solution of the whole computational domain. The unsteady Navier-Stokes equations were solved using the Smagorinsky-Lilly model, which simulates sub-grid scale effects and PISO algorithm to solve the coupling of velocity and pressure [18]. The momentum equation was calculated using the second-order central difference scheme. The second-order implicit scheme was used as time progressed, and the time-step of the unsteady calculation was

4.1 Measurement of integral characteristics Fig. 6 shows the schematic of the test rig and its component list. The section geometry of the test rig is rectilinear, and its size is 880 mm×880 mm. The flow direction is from left to right in this figure, and the test fan is set at the inlet part of the test rig. The flow rate was varied by controlling the rotating speed of the booster fan (#8, in Fig. 6) because the test rig does not have the throttle system to control the flow rate. The static pressure-rise was obtained from the pressure measured at the pressure tap (#2, in Fig. 6), and the flow rate was calculated from the difference between the upstream and downstream pressures of the nozzle (#7, in Fig. 6). The sample fan of airfoil A was tested to verify the correctness of the numerical calculation methods. The sample fan was set at the inlet part of the test rig, and its rotating speed was constant at 3000 r/m. The static pressure rise and the flow rate were varied by controlling the rotating speed of the booster fan, and the P-Q performance curve of the test fan was obtained. The results are obtained by comparing the P-Q performance curves obtained from the experiment and from numerical calculation. Both curves are roughly the same with each other, as shown in Fig. 7. Thus, the numerical calculation methods are feasible.

1680

L. Zhang et al. / Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685

(a) P-Q performance curve

Fig. 7. Two P-Q performance curves.

Fig. 8. Schematic view of the noise test rig. (b) η-Q performance curve Fig. 9. Integral characteristics of the fans.

4.2 Measurement of noise The international standard, ISO10302 [15], is a standard method for measuring the noise emission of small fans. The noise test was performed in a hemi-anechoic chamber, and the spatial dimensions of the chamber are 3.74 m × 3.34 m × 3.15 m. The background noise levels in the test chamber must be as low as possible, and the value should be 25 dB. Fig. 8 shows the schematic of the noise test rig. Test fans were fixed on the motor rotor and rotated along the motor rotor. The rotating speed of the test fans is constant at 3000 r/m. A sound level meter was placed 1 m away from the central axis of the impeller along the fan outlet to measure fan noise. Noise experiment was carried out on the sample fan of airfoil A. The experimental sound pressure level of aerodynamic noise is 36.6 dB compared with 34.8 dB based on numerical calculation. The deviation is less than 2 dB. Thus, the numerical calculation method of predicting fan aerodynamic noise is feasible. The experimental sound pressure level is larger than that of the numerical calculation because the impact of mechanical noise was ignored during the numerical calculation. A discrepancy always exists between experimental measurements and numerical simulation for any engineering problems. This discrepancy may be due to simulation error, experimental error, or both. In this study, the average results of multiple sampling were used to solve the uncertainty problems during the experimental measurement. A deviation less than 5% is still acceptable for engineering applications.

5. Results and discussion 5.1 Integral characteristics Fig. 9 shows the total pressure rise P and the total efficiency

η against flow rate Q of the fans of the three airfoils. From Fig. 9, the total pressures of the three airfoils were relatively similar with the whole flow rate range. The performance curves of airfoils A and C have concave features, which are characteristic of unstable conditions. However, the performance curve of airfoil B has a smooth feature. For flow rates Q < 0.006 kg/s, the total efficiencies of the three airfoils are almost similar to each other with the same flow rate. When the flow rate was increased to Q > 0.006 kg/s, the total efficiencies of airfoils A and C are almost similar to each other with the same flow rate. On the other hand, the total efficiency of airfoil B is significantly greater compared with those of the other airfoils. The highest efficiency of airfoil B is greater by 10% than those of the other airfoils. When the flow rate was increased, the efficiency deviation between airfoil B and the other airfoils became wider. 5.2 Total pressure loss coefficient The total pressure loss coefficient C pt is defined as C pt =

Pinlet −Poutlet 0.5ρ Vinlet 2

(3)

L. Zhang et al. / Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685

Fig. 10. Total pressure loss distribution with blade height.

where Pinlet is the relative static pressure of the fan inlet, Poutlet is the relative static pressure of the fan outlet, ρ is the fluid density, and Vinlet is the average relative velocity of the fan inlet. Fig. 10 shows the total pressure loss distribution of the three airfoils along the direction of blade height from the hub surface to the casing surface at the same flow rate (Q = 0.011 kg/s). The ordinate r/R indicates the relative distance close to the hub surface along the direction of blade height, and the abscissa represents the total pressure loss coefficient. Based on the figure, the curve of airfoil A shows an "M" shape, and the peak positions of the total pressure loss coefficient are distributed at 50% and 95% of the blade height. However, the curves of the other airfoils show an "N" shape, and the peak position of the total pressure loss coefficient is distributes at 75% of the blade height. The peak positions of airfoils B and C move to the center of the blade height compared with that of airfoil A. Thus, airflow conditions near the blade root and the blade tip are improved. In summary, airfoil C has the maximum total pressure loss, whereas airfoil B has the minimum total pressure loss among the three airfoils. This result is consistent with the total efficiency trend at the same flow rate (Q = 0.011 kg/s) in Fig. 9(b).

1681

(a) Trailing edge region

(b) Tip clearance region

5.3 Noise characteristics Fan noise mainly consists of vibration-induced noise and aerodynamic noise. In this paper, only the aerodynamic noise was considered. Aerodynamic noise is concerned with sound generated by aerodynamic forces or blade motion. The monitoring points of the three airfoils were set up at the trailing edge region, the tip clearance region, and the far-field region 1 m away from the fan hub to explore the influence of irregular blade designs on aerodynamic noise. Fig. 11 shows the sound pressure level of the three monitoring points of the three airfoils. The sound pressure levels of the aerodynamic noise of airfoils B and C are smaller than that of airfoil A within the whole frequency range. The deviation exceeds 10 dB within the frequency range of 4500 Hz-7000 Hz near the blade trailing edge region. Near the blade tip

(c) Far-field region Fig. 11. Sound pressure level of the three monitoring points.

clearance region, the sound pressure level of the aerodynamic noise of airfoil B is greater than that of other airfoils within the whole frequency range. Near the far-field region, the sound pressure levels of the aerodynamic noise of airfoils B and C are greater than that of airfoil A within the frequency range below 4500 Hz, whereas the opposite result was obtained within the frequency range above 4500 Hz where the sound pressure level of airfoil B is the lowest among the three airfoils. By comparing Fig. 11(b) with Fig. 11(c), the sound pressure level trends of aerodynamic noise of the three airfoils are

1682

L. Zhang et al. / Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685

low band and in the frequency range of 3000 Hz-18000 Hz in the high band. The aerodynamic noise of the low band has a discrete spectrum feature, which is characteristic of discrete noise, because the power spectral density represents the main indicators of the intensity of aerodynamic noise power. A gradual transition in the continuous spectrum occurs with increasing frequency, which is characteristic of broadband noise. The aerodynamic noise of the fan in the near-field region can be mainly classified as discrete noise, whereas that in the far-field region can be mainly classified as broadband noise. (a) Trailing edge region

(b) Tip clearance region

(c) Far-field region Fig. 12. Power spectral density of the three monitoring points.

exactly the same when the frequency is more than 4500 Hz and less than 11000 Hz near the blade tip clearance region and the far-field region. This result indicates a noise relationship between the blade tip clearance region and the far-field region. Fig. 12 shows the power spectral density of three monitoring points of the three airfoils. The power spectral densities of the three airfoils are mainly distributed in the blade passing frequency (BPF) and its harmonics are near the blade trailing edge region as well as the blade tip clearance region. Near the far-field region, the power spectral densities of the three airfoils are mainly distributed in BPF and its harmonics are in the

5.4 Mechanism analysis of the irregular airfoils The internal flow characteristics of a small axial flow fan are the key factors that affect the static characteristics and noise. Thus, analyzing the internal flow characteristics of a fan is a feasible method of determining the mechanism of the irregular designs. Fig. 13 shows the velocity streamlines and the axial velocity contours of the meridional plane (X = 0 mm, as shown in Fig. 14) of the three airfoils when the flow rate is 0.011 kg/s. From Fig. 13, the trunk stream regions of airfoil B are the widest and should be the least affected by the vortex. The trunk stream regions of airfoil A are relatively the narrowest. The trunk stream regions are associated with fan efficiency, which means that a wider trunk stream region results in a higher fan efficiency. Thus, the fan efficiency of airfoil B is the maximum, whereas the fan efficiency of airfoil A is the minimum. This trend of fan efficiency is consistent with that of Fig. 9(b). From Fig. 13(b), airfoil B changes the turbulent structure near the wall and squeezes the sphere of influence of the vortex. Thus, the vortex has to be mainly concentrated in the blade tip region of the leading edge of the fan. From Fig. 13(c), airfoil C can change the vortex structure of the blade trailing edge region and can scatter the vortex of the trailing edge into a number of small eddies. Fig. 15 shows static pressure contour distribution of the blade suction surface and blade pressure surface of the three airfoils when the flow rate is 0.011 kg/s. The lower pressure region in the suction surface and the higher pressure region in the pressure surface of airfoil A are significantly greater than those of other airfoils. The blade pressure difference of airfoil A is the maximum, whereas that of airfoil B is the minimum. The scopes of the tip leakage vortex of airfoil A are wider than those of the other airfoils because the pressure gradient between the pressure surface and the suction surface of the blade tip region is the main factor that forms the tip leakage vortex. The above trend of the scopes of the tip leakage vortex is consistent with that of Fig. 13. Fig. 16 shows the vorticity contour distributions of the revolution surface (R = 24 mm, as shown in Fig. 17) of the three airfoils when the flow rate is 0.011 kg/s. The vorticities of vortex shedding off airfoils A and B change along the axial direction, whereas that of airfoil C changes along the circumferen-

1683

L. Zhang et al. / Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685

Fig. 14. Geometrical position of the meridional plane X = 0.

(a) Airfoil A

(a)

(b) Airfoil B

(c)

(e)

(b)

(d)

(f)

Fig. 15. Static pressure contour distribution of the blade suction surface and the pressure surface of the three airfoils: (a) Suction surface of airfoil A; (b) Pressure surface of airfoil A; (c) Suction surface of airfoil B; (d) Pressure surface of airfoil B; (e) Suction surface of airfoil C; (f) Pressure surface of airfoil C. (c) Airfoil C Fig. 13. Velocity streamlines and axial velocity contours of the meridional plane.

tial direction. Airfoil C changes the flow direction of vortex shedding, which reduces the influence of vortex shedding off in the far-field region. Moreover, the vorticity gradient variation near the blade trailing edge of airfoil B is much smoother than that of airfoil A. Based on the vortex-sound theory, a

smoother vorticity gradient variation indicates a smaller aerodynamic noise. Thus, the aerodynamic noise of airfoil C is the minimum, whereas that of airfoil A is the maximum. This result is consistent with the trend in Fig. 11(a).

6. Conclusions This paper introduced irregular airfoils on small axial flow

1684

L. Zhang et al. / Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685

(a) Airfoil A

(b) Airfoil B

(c) Airfoil C

fans. With the help of numerical simulation and experimental testing, the effects of irregular airfoils on the aerodynamic performance and the internal flow field were determined, and the mechanism of the irregular airfoils of the fans were discussed. The following conclusions are drawn: (1) The total pressures of the three airfoils are relatively similar to each other over the whole flow rate range. The performance curves of airfoils A and C have concave features, which are characteristic of unstable conditions. However, the performance curve of airfoil B has a smooth feature. For flow rates Q < 0.006 kg/s, the total efficiencies of the three airfoils are almost similar to each other with the same flow rate. When the flow rate was increased to Q > 0.006 kg/s, the total efficiencies of airfoils A and C are almost similar to each other with the same flow rate. The total efficiency of airfoil B is significantly greater than that of the other airfoils. The highest efficiency of airfoil B is greater by 10% than that of other airfoils. When the flow rate was increased, the efficiency deviation between airfoil B and the other airfoils became wider. (2) Near the far-field region, the sound pressure levels of the aerodynamic noise of airfoils B and C are greater than that of airfoil A within the frequency range below 4500 Hz, whereas the opposite result was obtained within the frequency range above 4500 Hz. the sound pressure level of airfoil B is the minimum. (3) The aerodynamic noise of the low band has a discrete spectrum feature, which is characteristic of discrete noise. A gradual transition in the continuous spectrum occurs with increasing frequency, which is characteristic of broadband noise. The aerodynamic noise of the near-field region of the fan can be mainly classified as discrete noise, whereas that of the far-field region can be mainly classified as broadband noise. (4) Airfoil B changes the turbulent structure near the wall and squeezes the sphere of influence of the vortex. Thus, the vortex has to be concentrated mainly in the lade tip region of the leading edge of the fan. Moreover, airfoil C changes the flow direction of vortex shedding, which reduces the influence of vortex shedding on the far-field region.

Fig. 16. Vorticity contour distributions of the revolution surface R = 24 mm.

Acknowledgment This work was supported by grants from Zhejiang Science Technology Plan Project (2011C16038), the National Natural Science Foundation of China (51006090), and Zhejiang Science Technology Project (2011C11073).

Nomenclature------------------------------------------------------------------------

Fig. 17. Geometrical position of the revolution surface R = 24 mm.

P Q η N-S FW-H LES

: Total pressure rise : Flow rate : Total efficiency : Navier-Stokes : Fowcs-Williams and Hawkings : Large eddy simulation

L. Zhang et al. / Journal of Mechanical Science and Technology 27 (6) (2013) 1677~1685

References [1] W. R. Kundert, Fan speed control for noise control, Proceedings of noise-con (G. C. Maling, ed.), 85 (1985) 127132. [2] W. Neise and G. H. Koopman, Active source cancellation of the blade tone fundamental and harmonics in centrifugal fans, Proceeding of inter-noise (M. Bockhoff, ed.), 88 (1988) 801804. [3] D. A. Quinlan, Active control of tones radiated from small axial flow fans, Noise control engineering journal, 32 (1992) 95-101. [4] G. C. Lauchle, J. R. Macgillivray and D. C. Swanson, Active control of axial-flow fan noise, Journal of the acoustical society of America, 101 (1997) 341-349. [5] M. G. Beiler and T. H. Carolus, Computation and measurement of the flow in axial flow fans with skewed blades, Journal of Turbomachinery Transactions of ASME, 121 (1999) 59-66. [6] T. Sasaki and F. Breugelmans, Comparison of sweep and dihedral effects on compressor cascade performance, J. of Turbomachinery, 120 (1998) 454-464. [7] M. G. Beiler and T. H. Carolus, Computation and measurement of the flow in axial flow fans with skewed blades, ASME Journal of Turbomachinery, 121 (1999) 59-66. [8] S. M. Sun et al., Experimental research on noise reduction of bionic axial fan blade and mechanism analysis, Journal of Jilin University (Engineering and Technology Edition), 39 (2009) 382-387. [9] L. Q. Ren, S. M. Sun and C. Y. Xu, Noise reduction mechanism of non-smooth leading edge of owl wing, Journal of Jilin University (Engineering and Technology Edition), 38 (2008) 126-131. [10] B. I. Epureanu, K. C. Hall and E. H. Dowell, Reducedorder models of unsteady viscous flows in turbomachinery using viscous-inviscid coupling, Journal of Fluids and Structures, 15 (2001) 255-273. [11] B. You et al., Numerical analysis and experimental

1685

research on the internal flow characteristic of axial flow fans with dimples on suction surface of blades, Journal of Engineering Thermophysics, 29 (2008) 68-72. [12] E. Matjaz et al., Numerical and experimental investigation of axial fan with trailing edge self-induced blowing, Forsch Ingenieurwes, 73 (2009) 129-138. [13] I. T. O. Shinichiro, Aerodynamic influence of leading edge serrations on an airfoil in a low Reynolds number, Journal of Biomechanical Science and Engineering, 4 (2009) 117-123. [14] B. You et al., Numerical analysis and experimental research on the internal flow characteristic of axial flow fan with tooth shaped trailing edge, Journal of Engineering Thermophysics, 28 (2007) 592-594. [15] X. J. Lu, Experimental study on noise control of a centrifugal blower using tooth shape vane, Transactions of the Chinese society of agricultural machinery, 32 (2001) 86-88. [16] H. Y. Qian et al., Aerodynamic performance study on small axial fan with tooth shaped trailing edge, Journal of zhejiang Sci-tech university, 29 (2012) 374-377. [17] K. Chen et al., Aerodynamic noise reduction of small axial fan using hush characteristics of eagle owl feather, Journal of Jilin University (Engineering and Technology Edition), 42 (2012) 79-84. [18] G. X. Cui, C. X. Xu and Z. S. Zhang, Progress in large eddy simulation of turbulent flows, Acta Aerodynamica Sinica, 22 (2008) 121-129. [19] Q. G. Chen et al., Large eddy simulation of the discrete noise in an axial-flow fan, Fluid Machinery, 38 (2010) 18-22.

Li Zhang received his M.S. from Zhejiang University, China in 2006. His research interests include fan noise reduction and performance optimization. In 2009, he began studying for his doctorate degree in Zhejiang Sci-tech University.