Effect of a Gurney flap on the performance of an axial fan

Chen Liu, Haijun Xie, Jun Xu, Chen Jian*, Ren Dai 2 School of Energy and Power Engineering, University of Shanghai for Science and Technology 3 E-mail: 09900589r@connect.polyu.hk 4 Abstract: 5 The Gurney flap (GF) is a miniature lift-enhancement device and is usually mounted at the trailing 6 edge of an airfoil. The GF and has been successfully applied to isolated airfoils, multi-element 7 airfoils, and aircraft wings as well as helicopter rotor due to its attractive features of simplicity, cost8 effective and separation control. The GF also has aroused the attention of researchers in the 9 turbomachinery industry. However, limited studies are currently available on the application of a 10 GF to an axial fan. 11 Hence, in this paper, we conduct a wind tunnel and computational fluid dynamic (CFD) 12 investigation on an axial fan profiled with a NACA 65-(12)10 airfoil to evaluate the effect of the GF 13 on the performance of the fan. We also present the detailed flow features of the fan with and 14 without the GF after validating the simulation results with the experimental results. The 15 experimental results show that as the GF is installed higher on the fan blade, it can produce a higher 16 total pressure rise accompanied with a greater loss of efficiency. The installation of the GF also 17 enlarges the work capacity of the fan. Detailed flow field analysis, including the surface pressure 18 distribution, vorticity distribution at the trailing edge and streamline distribution of the fan, is 19 carried out to understand the mechanisms of the effect of the GF on the performance of the fan. 20


22
The Gurney flap (GF), invented by a race car driver Dan Gurney in 1960, is a small lift-enhancement 23 device and is usually mounted at the trailing edge of an airfoil. A Gurney flap with a height of 2% 24 of the airfoil chord (2%C) can increase the maximum lift coefficient by 21% and lift-to-drag ratio by 25 35% on a NACA0012 airfoil [1]. The lift enhancement mechanism stated by Liebeck is that the The results showed that the increment of the lift coefficient decreased when the GF was moved 37 forward to the leading edge. The mounting angle was also studied by Traub L.W., et al [4]. They 38 found that a perpendicular GF generated the lowest lift-to-drag ratio. GFs at 45°and 60° produced 39 the maximum lift-to-drag ratio. Kobayashi T.,et al[5] came to a similar conclusion. The 40 configuration of the GF was investigated by Dan and Vijgen, P.M., et al [6, 7] using a sawtoothed 41 GF for an isolated airfoil. They indicated that the sawtoothed GF improved the lift coefficient and 42 reduced the drag coefficient. This was attributed to a reduction in the effective windward area. 43 A large number of researchers have used GFs on dual or multi-element airfoils or wings for aircraft 44 applications. Storms B.L, et al [8,9] utilized a 0.5% chord GF and a 1% chord GF on a NACA632-45 215(B) airfoil to increase the maximum lift coefficient by 10.3% when the deflection angle of the 46 slotted flap was 42 degrees. Myose R et al [10,11]mounted a 1% chord GF at the trailing edge of 47 the main and flap elements on a two-element GA(W)-2 airfoil. It was found that the "Flap GF" and 48 "Flap main GF" can enhance the lift coefficient, but the "main GF" reduces the lift coefficient 49 compared to a "clean" airfoil. An investigation of the effect of a GF on multi-element airfoils can 50 also be found in references [12]. Traub plain GFs with heights ranging from 0.01C to 0.05C as well as four 0.05C serrated GFs with different 55 serration heights ranging from 0.01C to 0.05 C on a cropped nonslender delta wing. They found 56 that a proper GF can provide a significant improvement in performance. Zhan and Ji Wang[15]also 57 tested a delta wing with GFs and apex flaps. The experimental results indicated that the lift 58 coefficient was improved for different angles of attack (AOA) for a delta wing with GFs and apex 59 flaps. Albertani, R [16] used a simple GF to improve the aerodynamic characteristics of a micro 60 aerial vehicle wing. With GFs, the maximum lift coefficient was increased by 26.5% in comparison 61 with clean wings. Additional applications of the GF on wings can be found in [17,18]. 62 Successful applications of GFs on an isolated airfoil, multi-element airfoil or wings as well as the 63 been few studies of the application of GFs to fans, especially with regard to axial fans. The 85 application of GFs seems to be a promising approach to improve the performance of axial fans. 86 Thus, the purpose of this paper is to investigate the performance of an axial fan with GFs by 87 experimental measurement. Flow features were interrogated by the CFD (Computational Fluid 88 Dynamics) code to comprehensively understand the physics involved in the GF. The pressure 89 coefficient, total pressure efficiency and static pressure efficiency of the clean fan and axial fans 90 with varied plate GFs were tested first. Then, the simulation results were compared with the tested 91 results to validate the computational strategies. In the end, we analyzed and visualized the flow 92 features of a fan with GFs. The latter work offers a theoretical foundation for further optimization 93 of axial fans. . The flow quality of the 99 plenum is guaranteed. Beyond a velocity of 10 m/s, the variation in the total pressure and static 100 pressure field is less than 0.1% and 0.3%, respectively, while the maximum deflection of the flow 101 angle is below ±0.2 degree, and the turbulence intensity is less than 0.5% at the test section. 102 The flow plenum is divided into four parts (long entrance duct, diffusion nozzle, plenum and 103 exhaust duct). The lengths of these four parts are approximately 3580 mm, 587 mm, 1600 mm and 104 6 370 mm, respectively. The diameter of the long duct is approximately 498 mm, and the diameter 105 of the inlet chamber is approximately 1240 mm. Upstream of the long entrance duct, there is a 106 flow nozzle, where several additional taped holes are mounted. Circumferential tapes are used to 107 measure the static pressure difference across the flow nozzle to calculate the volumetric flowrate 108 More information about these axial fans can be found in Table 1. 138 We add the GF directly to the trailing edge of the blade of the prototype fan to form an axial fan 140 with a GF. Three flaps with different heights 1.25%, 2.5% and 5% of chord length are studied, which 141 means that the height of the GFs are only 0.51 mm, 1.03 mm and 2.05 mm. This is too small of a 142 dimension to be fabricated by a CNC. Thus, during testing, a special type of tape was adopted to 143 form the GFs. Figure 3 presents the real models of axial fans with and without GFs. The special 144 black tape has different thicknesses, and we can choose an appropriate thickness to form the GFs.   Table 2 presents detailed   160 information about the data acquisition devices, including the name, range and accuracy. 161

Data reduction 163
The purpose of fan testing is to obtain a performance map of the fan, including the pressure rise, 164 efficiency, and shaft power as a function of the flowrate or rotational speed. Through the above 165 data acquisition system, we can easily obtain values for a P , a T , h , 2 p , 6 p , in t P , , N and e P .

166
The static pressure through the inlet nozzle is given by 167 The mass flow rate can be calculated by the following formula using elementary mass conservation 169 and Bernoulli's law 170 The total pressure rise of the fan can be expressed in this form: 177 The static pressure rise of the fan can be expressed in the form: 186 The total pressure rise coefficient (TPRC) ,..., , 2 1 , they have affected the indirectly 207 measured values, so: 208 Comprehensive error can be expressed as: 210

217
The error of the gas constant can be calculated by the following formula: 218 The error of the atmospheric density can be calculated by the following formula: 222 The error of the volume flow can be calculated by the following formula: 226

231
The error of the exhaust airflow density can be calculated by the following formula: 232 The error of the efficiency of the fan can be calculated by the following formula: The final results are shown in Table 3. 244  to achieve a high quality mesh. Popular ICEM software is adopted to generate the unstructured 254 tetrahedral mesh. Boundary layers are paved on the solid wall surfaces of the blades, hub and 255 shroud to ensure that y-plus is less than 1 to resolve the near-wall flow properly. The mesh size is 256 also refined in the areas that are close to the axial fan to properly adapt to the large gradient 257 variation of the pressure and velocity. The detailed mesh around the fan blade is shown in Figure  258     The finite volume method is used to discretize the governing equations. The second-order upwind 287 interpolation schemes are applied for the pressure and momentum equation, respectively. The 288 SIMPLE algorithm is adopted to handle the pressure-velocity coupling. The frozen rotor method is 289 utilized to address the data transform between the rotational and stationary zone interface. A 290 convergence criterion is prescribed to obtain a periodic variation of the torque of the fan as well 291 as a maximum residual less than 1.0x10 -5 . 292 indicates that a higher GF produces a higher total pressure and maximum operating flow range. In 299 other words, a fan with a GF has a large operating flow range. The higher total pressure is also 300 related to the larger operation capability. 301 302 Figure 5 Effects of GF height on the total pressure rise coefficient 303 Figure 6 shows the effects of the height of the GF on the total pressure efficiency ( t  ) at different 304 flow coefficients. It is found that the total pressure efficiencies of fans with GFs are always lower 305 than that of the original. The difference between the total pressure efficiency curves is not obvious 306 in the flow coefficient range from 0.14 to 0.27 for fans with 1.25% and 2.5% chord GFs. As the flow 307 coefficient increases to 0.39, the difference among the four curves is increased. The total pressure 308 efficiencies at 30 . 0  q for 1.25%, 2.5%, 5% chord GFs are 1.76%, 4.6%, and 6.7% lower than 309 that of the original fan, respectively. All of the above observations indicate that GFs with 1.25% and 310 2.5% chords have very small effects on the total pressure efficiency at low flow coefficients. As the 311 flow coefficient reaches a certain value, the influence of the height of the GF on the total pressure 312 efficiency is remarkable. 313 1.04%, 1.65%, and 3.05% lower than that of the original fan, respectively. One possible reason for 321 the efficiency reduction may be the vortex formed by the GF at the trailing edge of the airfoil, 322 producing the loss. Another reason may be that the increased pressure difference between two 323 blade surfaces leads to an increase in clearance leakage. However, when the flow coefficient 324 becomes greater than 0.27, the opposite behavior is observed, in that the efficiency difference 325 becomes negligible with an increase in the height of the GF. Similar observations are found in the 326 studies of Manish et al [28]. 327

Results and Discussion
In some application circumstances of cooling, ventilation, vacuuming, dust removal and inflating, 328

345
The following numerical simulation was carried out under the condition that the mass flow q is 346 0.028 kg/s. Figure 9 shows the distribution of vorticity, which is generated by the Q-criterion at the 347 trailing edge of the airfoil, colored by the static pressure. It is found that the airflow over the 348 original fan is separated from the trailing edge on the suction surface and that a shedding vortex is 349 generated downstream of the trailing edge. When GFs with different heights are added at the 350 trailing edge of the airfoil, it is observed that a low-pressure vortex area is formed to reduce the 351 adverse pressure gradient in the flow channel and accelerate the flow of air. This is a possible 352 reason for the enhancement of the adsorption force at the trailing edge of the suction surface. 353 Thus, the blade load is improved significantly as well as the outlet total pressure.  Figure 11 shows the variation in the total pressure coefficient along the flow direction at the blade 376 midspan. A higher total pressure coefficient is obtained with a higher flap. 33.33% more total 377 pressure coefficient can be obtained with the GF height of 5%C than the original fan. Since a GF is 378 so limited in size, an obvious difference in the total pressure coefficient between fans with and 379 without GFs can only be found on the near airfoil and trailing edge of the blade. 380 381 Figure 11 Midspan axial variation of the total pressure rise. 382 Figure 12 depicts the streamline distribution on the S1 stream flow surface. It is observed that 383 separation occurs earlier on the original fan than on a fan with a GF. This indicates that the flow 384 separation at the suction surface will be suppressed when the GF is installed. 385 angle. Therefore, the GFs can enhance the work capability of fans. We know from the previous 403 sections that the GF is able to increase the pressure difference between the suction surface and 404 pressure surface, which produces a stronger blade tip leakage vortex, tending to move from the 405 pressure surface to the suction surface. Its mixing with the main flow is intensified with the height 406 of the installed GF. As shown in Figure 14