Vorte Sheing on Combine Boies at Incience to a Uniform Air Stream T. Yavuz, Y. E. Akansu, M. Sarıo lu, an M. Özmert. Ba kent Universit, : Nige Universit,, : Karaeniz Technical Universit,Turke Abstract Vorte-sheing phenomenon of the flow aroun combine two boies having various geometries an sizes has been investigate eperimentall in the Renols number range between 4.110 3 an 1.7510 4. To see the effect of the rotation of the boies on the vorte sheing, the combine boies were rotate from 0 to 180. The combine moels have a cross section composing of a main circular cliner an an attache circular or square cliner. Results have shown that s for two cases were change consierabl with the angle of incience, while it was foun to be largel inepenent of Renols number at 150. Characteristics of the vorte formation region an location of flow attachments, reattachments, an separations were observe b means of the flow visualizations. epening on the inclination angle the effects of flow attachment, separation an reattachment on vorte-sheing phenomenon have been iscusse. Kewors Bluff bo, vorte sheing, flow separation, flow reattachment. F I. INTROUCTION LOWS aroun circular cliner an square prism as basic geometries, have been investigate b man workers for a long time. Such bluff boies often applie in the form of groups in man cases of engineering practices. In these cases, flow characteristics of such objects in contact or in close proimit show major ifferences, because of the interactions with each other, accoring to those of single applications. Therefore, numerous investigators have been intereste in flow past two boies which have same or ifferent shapes. Fleck [1] investigate the flow past a combine bo, compose of a circular cliner an a rectangular prism, in the Renols number range from 110 4 to 510 4. It was foun that the observe is inepenent of Renols number for the range of angle of attack -90 45. In effect, the sharp corners of the rectangular portion ominate the vorte sheing mechanism for the majorit of the range of the angle of attack. The onl range where there is Renols number epenence is 45 90, when the rectangular bo is largel in the wake of the circular bo. In this case, the presence of the rectangular bo on the ownstream sie : Corresponing autor of the circular cliner consierabl affects the location of the separation point. Wei an Chang [2] stuie flow characteristic of wake an base-blee flow ownstream of two bluff boies, with ifferent geometries, arrange sie b sie. The two-bo arrangements are comprise of flat plate-square cliner, flat plate-circular cliner, an square cliner-circular cliner. In this stu, the use two sets of moels, these being bluff boies with the same cross sectional imension but with ifferent vorte sheing frequencies, an bluff boies with ifferent cross sectional imension but with the same vorte sheing frequencies. It is suggeste that when the gap istance is small, the vorte sheing frequenc ownstream of the two bo arrangement is about half of the average of the sheing frequencies that correspon to each single bluff bo. Luo an Gan [3] investigate the structure associate with flow past two tanem circular cliners with a iameter ratio of 0.33 an with the smaller cliner upstream. Just like the case of flow past two equal size cliners, a critical spacing was foun to eist. For spacing less than the critical value, share laers that separate from the upstream cliner reattach onto the ownstream cliner whereas for spacing larger than the critical value both cliners she vortices. In the stu of Igarashi [4], eperimental investigations on the characteristics of a flow aroun two circular cliners of ifferent iameters with the ratio 2 / 1 =0.68 arrange in tanem were carrie out. The flow patterns var with an increasing of the spacing of the aes of the cliners in the same manner as in the case of equal iameters. The pattern in which the spacing of the aes of the cliners was the smallest (this case is closest to our present case) is a complete separation tpe in which the separate shear laer from the first cliner oes not reattach onto the secon one. Tsutsui et al. [5] stuie eperimentall an numericall the behavior of an interactive flow aroun two circular cliners with ifferent iameters (/=0.45) at close proimit. In particular, the reattachment that the separate share laer from the main cliner reattaches on the rear surface b Coana effect was mentione. In the stu of Gu an Sun [6], the interference between two ientical circular cliners arrange in staggere configurations has been investigate at high subcritical Renols numbers. In general, three ifferent pressure 673
istribution patterns on the ownstream cliner an two switching processes were observe for the win angle varing from 0 (in tanem) to 90 (in sie b sie). The corresponing flow patterns were classifie as wake, share laer an neighbourhoo interference respectivel. For ifferent forebo sections, vorte sheings from bluff boies with emphasis on fining the effects of afterbo shape were eperimentall investigate b Nakamura [7]. It was foun that the of a bluff bo with afterbo ecreases initiall with increasing sie ratio in which one of the most important shape parameters for controlling the is the ratio of afterbo length to cross-flow imension of the bluff bo. This is in sharp contrast to the base suction that is sensitivel epenent of afterbo shape. Nakamura [8] also eperimentall stuie the effect of the etene splitter plates on the bluff boies with forebo shapes. These forebo shapes eamine inclue a circular shape, a semi-circular section with an without a rectangular block an a normal flat plate. It was shown that vorte sheing from bluff boies with etene splitter plates is characterize b the impinging-shear-laer instabilit, where the in terms of splitter plate length increases with increasing splitter plate length. Akansu et al. [9] investigate eperimentall the behavior of a stationar circular cliner with an attache plate, uner conitions where the entire cliner-plate bo was to rotate about the cliner ais, in the Renols number of 210 4. The results inicate that the sheing frequenc was nearl constant in the range of the plate angle of 50 120 an as farther increasing the angle from 120 to 160, it strikingl increases an then again ecreases at the angles bigger than 160 eg. The plate also causes important changes in pressures on the surface of the cliner epening on the flow separation an reattachment as increasing the inclination angle. A similar stu for a square prism with an attache plate was mae b Sarioglu et al. [10]. Their results inicate that the base on, the sie length of the square cliner, has a strong peak at = 12 an rag coefficient of the square cliner has minimum an maimum values at approimatel = 20 an 80 respectivel. Base on the results summarize above, it can be conclue that flow aroun two ajacent bluff boies epens strongl on the geometries an the angle of incience. Therefore, the present stu focuse on the investigation of the flow aroun combine bluff boies with ifferent cross-sectional geometries at incience. Two combine moels use as a main circular cliner (=9.5 mm) with an attache small cliner having circular cliner (=6 mm) or square prism (=6 mm). II. EXPERIMENTAL APPARATUS AN PROCEURE The eperiments were conucte in the test section of a blower tpe open jet TE 44 subsonic win tunnel having working cross section of 457 mm 457 mm. Bounar laer correction is achieve b corner fillets etening the length of the contraction cone an the working section, proportione to ensure a uniform longituinal pressure in the working section. At the maimum tunnel spee of about 30 m/s, the free stream turbulence intensit was about 0.5%; the turbulence intensit was higher at low tunnel spees, about 1.5% at 4 m/s, which is the lowest spee in the tunnel. The Renols number base on the iameter of the main cliner,, was range between 4.110 3 an 1.7510 4. The tunnel an test section are shown in Fig. 1. VALVE VENTILATOR ELECTRIC MOTOR 1200 TEST REGION STATIC PRESSURE TAPS Fig. 1 Blower tpe open jet subsonic win tunnel an test section. The test moels consiste of two configurations are shown in Fig. 2. As shown in the figure, the upstream cliners having circular an square cross sections were combine with the main circular cliner in a line contact at the center ais. The sie length of the square cliner (=6 mm) an the iameter of the main circular cliner (=9.5 mm) were chosen to ensure the vorte sheing frequencies of both square an circular cliners to be same when the sta isolate in the flow. All moels were mae from stainless steel an the square prism was machine sharp eges. The lengths of the boies are same with the with of the tunnel test section an the spanne the entire with of the test section. Flow Flow Combine Moel-I Combine Moel-II Fig. 2 Two ifferent combine moel geometries an coorinate sstem. The combine moels were centere on the mi-height of the test section an the were positione at an angle of incience to the free-stream flow irection an turne in the range of 0 180 with an increment of 3, with an accurac of 0.5. Maimum soli blockage ratios of the combine moels are, 3.48 % for the Moel-I at = 75 an 674
105, an 3.39 % for the Moel-II at = 90. As the blockage ratios are smaller than 6.0 %, no correction was mae for the blockage effects [11]. Minimum aspect ratios base on the projecte cross stream imension of the moels are 29.5 for the Moel-I an II. It can be sai that these aspect ratios are satisfactor for a bluff bo to be treate as two imensional [2]. Vorte sheings from the test moels were etecte b using a TSI IFA 100 moel constant temperature anemometer with two hot-film probes. The s for the vorte sheing from the moels were etermine from the frequenc analsis of the velocit fluctuations. The probes were locate at the position of / = 5 an 15, / = 2.5, where / an / are the ownstream an vertical istances respectivel, in the ownstream of the bo. The hot-film probes were calibrate using a TSI Moel 1125 calibrator. The frequenc response for the hot-film probes were foun to be about 20 khz b using a square wave test. The velocit measurements were carrie out using a computer controlle ata acquisition sstem. For velocit measurements at each measurement point, 4096 ata were acquire at a sampling rate of 4 khz using a low-pass filter setting of 2000 Hz. So the measuring time correspone to 1.024 s. TSI Thermal-Pro Software was use to acquire signals with a 12 bit A/ converter an to obtain the statistical an spectral analze of these measurement signals. Flow visualization eperiments were conucte b using a smoke-wire metho in the ifferent win tunnel for Renols number of 5.010 3. In these eperiments, three times bigger test moel imensions have been use to obtain a Renols number to be in the range of Renols number use in the velocit measurements. The eperimental uncertaint in the measurement of velocit was etermine to be less than 3%, whereas those of the an the frequenc spectra calculate from the eperimental measurements were etermine to be less than 3.3% an 0.5%, respectivel. III. EXPERIMENTAL RESULTS AN ISCUSSION The vorte sheings from the moels were eamine in the Renols number between 4.110 3 an 1.7510 4. A sample velocit spectra istribution obtaine for the Moel-I, at the Renols number 910 3 is given in Fig. 3. It is quite clear that, with increasing from 0, there is a sharpl ecrease in vorte sheing frequenc up to =24, then this ecrease in sheing frequenc slightl continues up to 48. After 84, it grauall increases up to 168 an then it ecreases again. Using vorte-sheing frequencies obtaine from spectral istributions, the s, St base on an St base on, calculate for the Moels I an II at the three Renols number 4.110 3, 9.010 3 an 1.510 4 are shown in Figs. 4 an 5. As shown in Fig. 4, at = 0 the value of St is about 0.25 an this value is above of that of the single circular cliner. This inicates that the shear laers separate from the leaing eges of the square cliner goes ver near main cliner of the Moel-I an causes to narrow wake than that of the single circular cliner. The ecreases sharpl from 0 to 15. Because, the shear laers separate from the corner A an B roll rear of the moel without a reattachments on the cliner an with a more wie wake. The increases grauall with the rotation of the Moel-I in the range between 15 69. This increase in St quietl epens on the reattachment of the separate shear laer from corner B on the face AB. As the angle approaches 69, this reattachment point closes to the corner A an, so it is seen a peak in the. After this angle, the separate shear laer from the corner B oes not reattach on the face AB anmore an causes a suen increase in the with of the wake an consequentl the ecreases. This phenomenon implies the eistence of a ecrease in the. This ecrease procees until about 90 in which the with of the wake reaches its maimum value. As the angle farther increases, the Strouhal number increases grauall up to 168 in which the Strouhal number shows a peak for the Renols numbers 4.110 3 an 910 3. Approaching this angle, a reattachment occurs again on the square cliner. After 168, there is a suen ecrease in ue to giving up this reattachment an also reucing the with of the wake. For Re=1.510 4, the peak in the occurs at 174. This is associate with the variation of the separation point on the circular cliner with the Renols number. As a matter of fact, in the stu of Fleck [1], changes with the Renols number strikingl especiall in the case where the circular cliner is upstream. Spectral ensit (arbitrar scale) 100 200 300 400 500 600 Frequenc, Hz 180 168 156 144 132 120 108 96 84 72 60 48 36 24 12 =0 Fig. 3 Spectra measure at / = 15, / = 2.5 in the wake of the Moel-I ( ). 675
0.3 0.2 FLOW A B 0.1 St = f/u St ' = f'/u Re = 4.110 3 Re = 4.110 3 0.0 0 30 60 90 120 150 180 C [ ] To eamine the flow fiel about the Moel-II arrangement, flow visualization eperiments were conucte b using a smoke-wire metho in the win tunnel for Re=5 103. In Fig. 6, the positions of flow attachment, separation, an reattachment on the two combine circular cliners ue to the inclination of the bo can be clearl seen. The variation of vorte formation region behin the boies can be also seen in this figure. When the flow pattern of the Moel-II at =0 is compare with the case of =180, there are consierable changes in the wake, an the istance require for vorte formation behin the boies varies. Narrow wake causes more small vortices an higher vorte sheing frequencies. Fig. 4 vs rotation angle, for the Moel-I. s for the Moel-II are presente in Fig. 5. In this case, as the smaller bo has circular shape, the Renols number effects are notice in the range of, 0 12, unlike the other moel. At = 0, the value of the Strouhal Number for the Re=4.110 3 is above of those for the Re=9.010 3 an 1.510 4. For the case of Re=4.110 3, at the beginning ( = 0 ), the shear laer separate from the small cliner reattaches on the ownstream cliner an this reattachment continues, with the small change in the location of the separation point on the big cliner, until = 12. After 12, a suen ecrease in the occurs b ening the reattachment. For the case of Re=1.510 4, there is no reattachment at the beginning. With increasing the angle of incience up to 12, the increases because of the reattachment occurring on the lower sie of the Moel-II. Here, also after = 12, the reattachments give up an the ecreases suenl. In the range of = 15-150, unlike the Moel-I having sharp eges, there is no peak in. After 150, the variation of the Strouhal number is similar to those of the Moel-I because of changing the position of the main cliner from ownstream to upstream, an consequentl the flow uner consieration is in the influence of the main circular cliner. 0.3 0.2 FLOW 0.1 Sr = f/u Sr ' = f'/u Re = 4.110 3 Re = 4.110 3 0.0 0 30 60 90 120 150 180 [ ] Fig. 5 vs incience angle for the Moel-II = 0 = 30 = 90 = 155 = 15 = 60 = 120 = 180 Fig. 6 Smoke-wire visualization of the flow aroun test Moel-II for Re = 5 10 3 Fig. 7 shows as a function of Renols number for various angles of incience consiere for the Moel-I. It can be sai that the was foun to be largel inepenent of Renols number up to 100. In the range of 160-175, has an unstable variation with Renols number in Re<7.510 3. This situation was also seen in Fig 4. 676
0.30 0.25 0.20 0.15 0.10 = 0 = 5 = 30 = 70 = 100 = 160 = 165 = 170 =1 75 it was foun to be largel inepenent of Renols number at 150 eg. For both of the moels, the increase at the angles of incience in which reattachments occur. B the en of reattachments, the with of the wake increases an ecreases suenl. The peak appeare in at about = 60 in the case of sharp ege square moel have not seen in the case of Moel- II which is compose of two circular cliners. Probe position : /=5; /=2 0.05 3.010 3 6.010 3 9.010 3 1.210 4 1.510 4 1.810 4 Renols number Fig. 7 istribution of of the Moel-I vs Renols number for ifferent angles of incience. A summarize classification of the flow regimes of the two combine test moels is shown in Fig. 8. The flow structures mentione above are presente schematicall in this figure. Here, flow attachments, separations an reattachments an their behaviors accoring to the angle of incience are shown. = 0 = 10 = 30 = 69 =9 0 = 120 = 168 = 180 (a) = 0 = 10 = 30 = 69 = 90 = 120 = 165 =180 (b) Fig. 7 Assume flow patterns aroun the combine moels for Re=9.010 4 a) Moel-I, b) Moel-II IV. CONCLUSION Vorte-sheing phenomenon in the flow aroun two ifferent combine moels at incience has been investigate eperimentall in the Renols number range of 4.110 3 an 1.7510 4. The following conclusions were obtaine: The s for two combine moels consiere were change strikingl with the angle of incience, whereas Smbol Quantit NOMENCLATURE iameter of the bigger circular cliner the projecte cross stream imension of the combine boies small imension of the upstream boies f vorte-sheing frequenc R Renols number base on, U/ St, f /U St, f /U U freestream velocit w the length of the big sie of the rectangular bo, streamwise an lateral coorinates inclination angle of the combine moels kinematic viscosit of flui ensit of air REFERENCES [1] B. A. Fleck, Strouhal Numbers for Flow past a Combine Circular- Rectangular Prism, Journal of Win Engineering an Inustrial Aeronamics, vol. 89, 2001, pp. 751-755. [2] Y. C. Wei an J. R. Chang, Wake an Base-Blee Flow ownstream of Bluff Boies with ifferent Geometr, Eperimental Thermal an Flui Science, vol. 26, 2002, pp. 39-52. [3] S. C. Luo an T. L. Gan, Flow past 2 Tanem Circular-Cliners of Unequal iameter, Aeronautical Journal, vol. 6, No. 953, 1992, pp. 105-114. [4] T. Igarashi, Characteristics of a Flow aroun Two Circular Cliners of ifferent iameters Arrange in Tanem, Bulletin of the JSME, vol. 25, No. 201, 1982, pp. 349-357. [5] T. Tsutsui, T. Igarashi, an K. Kamemoto, Interactive Flow aroun two Circular Cliners of ifferent iameters at Close Proimit. Eperiment an Numerical Analsis b Vorte Metho, Journal of Win Engineering an Inustrial Aeronamics, vol. 69, no. 71, 1997, pp. 279-291. [6] Z. Gu an T. Sun, On Interference between two Circular Cliners in Staggere Arrangement at High Subcritical Renols Numbers, Journal of Win Engineering an Inustrial Aeronamics, vol. 80, 1999, pp. 287-300. [7] Y. Nakamura, Vorte Sheing From Bluff Boies an a Universal Strouhal Number, Journal of Fluis an Structures, vol. 10, 1996, pp. 159-171. [8] Y. Nakamura, Vorte Sheing From Bluff Boies with Splitter Plates, Journal of Fluis an Structures, vol. 10, 1996, pp. 147-158. [9] Y. E. Akansu, M. Sarioglu, an T. Yavuz, Flow Aroun a Rotatable Circular Cliner Plate Bo at Subcritical Renols Numbers, AIAA Journal, vol. 42, no. 6, June 2004, pp. 1073-1080. [10] M. Sarioglu, Y. E. Akansu, an T. Yavuz, Flow Aroun a Rotatable Square Cliner Plate Bo, AIAA Journal, vol. 44, no. 5, Ma 2006, pp. 1065-1072. [11] G. S. West an C. J. Apelt, The Effects of Tunnel Blockage an Aspect Ratio on the Mean Flow Past a Circular Cliner with Renols Numbers between 10 4 an 10 5, J. Flui Mech., vol. 114, 1982, pp. 361-377. 677