A Study Of Aerodynamic Characteristics Of S-Shaped Savonious Rotor With Different Number Of Blades

Transcription

1 A Study Of Aerodynamc Characterstcs Of S-Shaped Savonous Rotor Wth Dfferent Number Of Blades Rumana Hossan Department of Physcal scence, Independent Unversty, Bangladesh Shaukat Ahmed Department of Industral and Producton Engneerng, BUET Abstract The S-shaped rotor s a modfcaton of the Savonus rotor wth no overlap. In ths paper the drag and torque coeffcents of statonary S-shaped two, three, four bladed rotor have been nvestgated by measurng the pressure dstrbuton on the blade surfaces for varous rotor angles. The experments have been carred out at Reynolds number 1.1*1 5 n a unform flow jet produced by an open crcut wnd tunnel. The measurements ndcate that drag force and so that the torque, vares wth rotor angle. The results facltate predctng the performance of these types of rotor under dynamc condtons. It s expermentally obtan that f the number of blade s ncreased there s no sgnfcant change n the tangental drag co-effcent but the startng torque ncreases drastcally due to the ncrease of pressure dfference and hence the normal force actng on the rotor blade. Keywords: S-shaped Savonus rotor, Drag coeffcent, Torque co-effcent 1. Introducton The wnd s an nexhaustble source of energy. In fact t can be regarded as an ndrect form of solar energy, the movements of the atmosphere resultng as they do from the ntermttent heatng effects of the sun's radaton on the ar tself and on the earth and the sea. The wnd s provded mechancal power for pumpng water and for mllng gran through many centures, and contnues to serve as the electrcal power source of many thousands of homesteads and farms n the rural areas. Throughout hstory people have harnessed the wnd. More than 5 years ago, the Egyptans used the wnd to sal shps on the Nle. Later, people bult the frst turbnes to grnd gran. These machnes looked lke paddle wheels and were used n Persa as early as B.C. By the 14th century, the Dutch had taken the lead n mprovng the desgn of wndmlls. They nvented propeller type blades and used wnd power to dran the marshes and lakes of the Rhone Rver delta. In Amerca, early European settlers used wndmlls to grnd wheat and corn, to pump water, and to cut wood at sawmlls. By the early twenteth century, small wndmlls were used for pumpng water and electrc power generaton n Europe, the Unted States, Afrca, and elsewhere. In addton to thousands of small wnd electrc generators, a few larger systems were bult n North Amerca and Europe. In the 197's, ncreases n the prce of ol and other fossl fuels helped wnd power return as an economcally vable, alternatve source of energy. Governments all over the world, especally n North Amerca and Europe, nsttuted research and development programs. These efforts led to the development of modern wnd turbnes, whch have dramatcally reduced the cost of generatng electrcty from wnd power. As a result, wnd energy contrbuted 3.8 bllon klowatt-hours of electrcty n 1991, enough to power more than 5, households. Blowng wnd spns the blades on a wnd turbne just lke a large toy pnwheel. The blades are attached to a hub that s mounted on a turnng shaft. The shaft goes through a gear transmsson box where the turnng speed s ncreased. The transmsson s attached to a hgh speed shaft whch turns a generator that makes electrcty. If the wnd gets too hgh, the turbne has a brake that wll keep the blades from turnng and beng damaged. Modern wnd turbnes are dvded nto two major categores: horzontal axs turbnes and vertcal axs turbnes. Horzontal axs turbnes are the most common turbne confguraton used today. They consst of a tall tower, atop whch sts a fan-lke rotor that faces nto or away from the wnd, the generator, the controller, and other components. 146

2 Most horzontal axs turbnes bult today are twoor three-bladed, although some have fewer or more blades. Vertcal axs turbnes fall nto two major categores: Savonus and Darreus. Nether turbne type s n wde use today. prmary attracton of the vertcal axs wnd turbne s the smplcty of ts manufacture compared to the horzontal axs wnd turbne. Most of the researchers have concluded research work on S- shaped rotor wth varous flow parameters. The Darreus turbne was nvented n France n the 19s. Often descrbed as lookng lke an eggbeater, ths vertcal axs turbne has vertcal blades that rotate nto and out of the wnd. Usng aerodynamc lft, these turbnes can capture more energy than drag devces. The Gromll and cycloturbne are varants on the Darreus turbne [4]. Frst nvented n Fnland, the Savonus turbne s S- shaped f vewed from above. Ths drag-type VAWT turns relatvely slowly, but yelds a hgh torque. It s useful for grndng gran, pumpng water, and many other tasks, but ts slow rotatonal speeds are not good for generatng electrcty [4]. The cost of energy from the wnd has dropped by 85% durng the last years. Incentves lke the federal producton tax credt and net meterng provsons avalable n some areas mprove the economcs of wnd energy. The rate the utlty pays to purchase wnd-generated electrcty, called the buy-back rate, often determnes the economc feasblty of a project. Wnd energy s often attractve when lfe-cycle costs are compared wth other generaton technologes. Wth utlty deregulaton, wnd energy may be mpacted by added costs for ancllary servces, the support servces needed to generate and delver electrcty.. Lterature Revew From the ancent perod people have been workng on varous types of wnd turbne to extract energy. However, expected developments have not been acheved n wnd drven machnes to cope wth the characterstcs of wnd turbnes. Conventonal machnes are beng used now-a-days even though they are not always sutable from the operatonal pont of vew. Wnd turbnes n ancent perod were of drag types only and ther effcences were not up to the mark. Arsng from the ncreasng practcal mportance of wnd turbne aerodynamcs, there have been, over the past few decades, and enormous ncrease n research works concernng laboratory smulatons, full scale measurements and more recently, numercal calculatons and theoretcal predctons of flows over a wde varety of Savonus rotors. Researchers n dfferent countres have been contrbutng greatly to the knowledge of analytcal predcton methods of wnd turbnes, but the major part of the reported works have been of fundamental nature nvolvng the flow over horzontal axs wnd turbne and vertcal axs Darreus rotor. The Islam et al. [5] nvestgated that the aerodynamc forces actng on a statonary S-shaped rotor and made an attempt to predct the dynamc performance from these forces. The work was done by measurng the pressure dstrbuton over the surfaces of the blades. The measurements were carred out n a unform flow jet produced by an open crcut wnd tunnel and at a constant wnd speed of 8.9 m/s whch corresponds to a Reynolds number of 1.13*1 5. The results ndcate that flow separates over the front and back surfaces of the blades and the pont of separaton depends on the rotor angle. The pressure dfference was observed between the convex and concave surfaces of each blade, whch n turn gave drag coeffcents. The drag and hence, the torque of each blade vares wth rotor angle and was found to be maxmum at =45 for the advancng blade and at =45 for the returnng blade. The maxmum net torque was obtaned at =45 and became negatve n the range of rotor angle between135 to 165. Islam et al. [6] analyzed the performance of S- shape rotor by placng a flat plate n front of the returnng blade. They found that the power coeffcent of the S-rotor was dependent on the Reynolds number. The value of Cp (power coeffcent) ncreased wth Reynolds number n the range of the Reynolds number studed and a maxmum of % of power coeffcent was ncreased by usng the deflectng plate whch occurred at a deflectng angle of 35 for b=.5d, where b=dstance between plate and rotor center and D=dameter of the rotor. There are several varatons of Savonus rotor. The effcency of a Savonus s only around 15 % but they are deal for many stuatons. They wll start spnnng wth very lttle wnd, produce a great amount of torque, and need no speed lmter. The Savonus rotor s desgned for low RPM applcatons, and wll never spn faster than the speed of the wnd. Some varatons are shown below, lookng down from the top of the turbne. 3. Detals of Non-Dmensonal Parameters In ths project, the local surface pressures on the blade surfaces of the rotor were measured relatve to the free upstream statc pressure prevalng n the test secton of the wnd tunnel. The data obtaned expermentally have been presented n terms of 147

3 non-dmensonal coeffcents. The equatons of coeffcent of pressure, normal and tangental drag co-effcent and torque coeffcent are brefly descrbed n ths secton. 3.1 Statc Aerodynamc Characterstcs In the followng two steps the procedure for calculatng the statc aerodynamc characterstcs s mentoned: Step1: From the expermental data pressure coeffcents both for concave and convex surfaces of blades are calculated for angle of rotaton rangng from to 34. Step: The flow pattern produced by the rotor blades s characterzed by hgh turbulence, unsteadness and flow separaton. The combned effect of these flow features n turn produces pressure dfferences between the concave and convex surfaces of the blades. These pressure dfferences result n aerodynamc forces and hence torques. These forces can be resolved nto components F n and F t along normal and tangental drectons of the rotor chord Pressure Coeffcent, C p In the experment, the pressure measurements were made at 8 tappng ponts on each blade for the angle of rotaton (α) at every nterval rangng from to 34 on both concave and convex surfaces. Usng these raw data pressures, coeffcent on convex and concave surfaces of the blades are measured by the followng equaton: C P P P 1/ U Where, P-P = Dfference between blade and atmospherc pressure. ρ = Densty of ar. U = Free stream velocty. C Fn ; C 1/ U R n Ft 1/ U t Where subscrpt ndcates the pressure tappng number. The forces F n and F t can be evaluated from the pressure dfferences as Fn P S sn ; Fn P S cos Where ΔP s the pressure dfference between the concave and convex surfaces of the blade at a partcular pressure tappng, ΔS s the chord length around the tappng, where a unform pressure s assumed R Fgure 1. Forces actng on rotor blades Torque Co-effcent The normal force s responsble for producng a torque on the rotor shaft and ths torque can be expressed for an element of the blade as T F n r 3.1. Drag Coeffcents These pressure dfferences between the concave and convex surfaces of the blades result naerodynamc forces and hence torques. These forces can be resolved nto components F n and F t along normal and tangental drectons of the rotor chord as shown n Fgure 1. The correspondng drag coeffcents can be expressed as Torque coeffcent can be wrtten as C q T 1/ U RR / Total drag coeffcent and torque coeffcent for an ndvdual plate at a partcular rotor angle, α are evaluated by ntegratng the coeffcents over the blade surfaces. Then 148

4 C ( ) C ; n n C ( ) C ; t 8 t water and 1475 rpm. A butterfly valve actuated by a screw thread mechansm has been placed behnd the fan and used to control the flow. A slencer has been ftted at the end of the flow controllng secton n order to reduce the nose of the system at the same tme a flow straghter system has been used n ths secton. C ( ) q Expermental Set-up and Procedure The objectves of ths project to nvestgate the effect of wnd on the S-shaped four, three, and two bladed Savonus rotor have been performed wth the help of a subsonc wnd tunnel, expermental set-up of the Savonus rotors and an nclned mult manometer. Pressure dstrbutons on the convex and concave surfaces of the blades of the S-shaped Savonus rotors placed normal to the approachng unform flow have been measured wth the help of the mult manometer. The followng sectons descrbe n detal about the expermental set-up and the technques adopted for the nvestgaton. 4.1 The Wnd Tunnel The open crcut subsonc wnd tunnel s 6.55m long wth a test secton of (49mm 49mm) cross secton. Fgure depcts the wnd tunnel used for the experment. The successve sectons of the wnd tunnel comprse of convergng mouth, Perspex secton, rectangular secton, fan secton (two rotary axal flow fans ), butterfly secton, slencer wth honey comb secton, an eddy breaker, dvergng secton, convergng secton, 61 mm long rectangular secton, 35 mm long flow straghter secton and 61 mm long rectangular ext secton. The central longtudnal axs of the wnd tunnel has been mantaned at a constant heght from the floor. The convergng mouth entry has been ncorporated nto the system for easy entry of ar nto the tunnel and to mantan unform flow nto the duct free from outsde dsturbances. The exposed surface of the mouth entry has been covered wth cloth for extra protecton aganst dust and for unform flow. The Perspex secton (35mm 35mm) s 61 mm long that may be used for the experment of flow around the crcular cylnder. A 9 mm long rectangular dvergng secton has been used to make the flow unform from rotaton. The nduced flow through the wnd tunnel has been produced by a two stage rotatng axal flow fan (Woods of Colchester Ltd., England, Type 38 JTE) of capacty m 3 /s at a head of 15.4 mm of C q The dvergng and convergng secton of the wnd tunnel s 46 mm long and made of 16 SWG black sheets. The angle of dvergence and convergence s 7, whch has been made wth a vew to mnmze expanson and contracton loss and to reduce the possblty of flow separaton. Other three square sectons have been used to make the flow straght and unform. Fgure. Schematc Dagram of Wnd Tunnel 1. Convergng Mouth Entry;. Perspex Secton; 3. Rectangular Dvergng Secton; 4. Fan Secton; 5. Butterfly Secton; 6. Slencer wth Honeycomb Secton; 7. Dvergng Secton ; 8. Convergng Secton; 9. Rectangular Secton; 1. Flow Straghter Secton; 11. Rectangular Ext Secton 4. Constructon of the S-shaped Savonus rotor The rotor blade desgn s geometrcally dentcal to that detaled by Lttler [1].The four, three, two bladed rotors are made up of four, three, two blades of chord length mm each and heght, H = 34.9 mm each. The blades are made of PVC materal. Each rotor s fxed on an ron frame by usng two sde shafts and two ball bearngs. The pressure measuremenst have been made at 8 tappng ponts on any two blades of each rotor. The tappng are made wth copper tubes that have been press ftted to the tappng holes. The tappng are located at the md plane on one sde of each blade, so that pressure at every on the blade surface can be measured. The pressure tappng have been connected to an nclned mult-manometer (manometrc flud s water) through PVC tubes. 4.3 Expermental procedure The flow velocty n the test secton was kept constant at 8.7 m/s. The Reynolds number (Re = UD/ν) based on rotor dameter, D = 19.5mm was The effect of the temperature was consdered n ths experment and the experment was carred at atmospherc temperature.e, at 149

5 Normal Drag Coeffcent Tangental Drag Coeffcent T=5C where the knematc vscosty of ar was m /s. The Savonus rotor wth frame was placed 5 mm upstream n front of the ext secton of the wnd tunnel. One blade of the rotor was fxed parallel to the free stream velocty.e, parallel to horzontal, whch was called the reference plane and from ths plane, angle of rotaton was measured. Then the frst blade was kept at angle of rotaton, exposng the convex surface n front of the free stream of ar, and other blades are kept at partcular angles. The rotor was made statc fxng one sde of the shaft of the rotor wth the anglefxng devce. Pressure on the convex surfaces of blades was measured at a partcular rotor angle α, keepng the rotor statc. Gradually pressure on the convex surfaces of blades was measured n ths process for every nterval. The same steps were repeated for the measurement of pressure on concave surfaces. At a partcular rotor angle α, the rotor blades had experenced forces due to pressure dfferences between the concave and convex surfaces, whch were resolved nto normal forces and tangental forces. 5. Results and Dscusson and after that t ncreases upto 16 o. From 16 to t s almost constant and then ncreases abruptly up to 8 and then t decreases drastcally. Mnmum negatve normal drag occurs at α= 1 and maxmum postve occurs at rotor angle 8. For the three bladed rotor the value of C n starts ncreasng from the and reaches ts peak at α=4 then t decreases wth fluctuaton as the α ncreases. In case of four bladed rotor the value of C n ncreases from α= to α=6 and t has the hghest postve value at ths pont. Then t decreases abruptly up to α=1 and after that t starts ncreasng agan up to α=18. From 18 to of rotor angle t s almost constant. After α= normal drag co-effcent decreases up to α=8 and at that pont t has the lowest negatve value. After α=8 t starts ncreasng agan up to 3. Overall the torque co-effcent shows a cyclcal pattern n whch decreases eventually for the four bladed rotor. Tangental drag coeffcents are plotted aganst dfferent rotor angles n Fgure 4. Angle of Rotaton Vs Total Tangental Drag Coeffcent for Two,Three and Four Bladed Rotor The pressure dstrbutons over the surfaces of the blades are measured at every 8 nterval of rotor angle for two, three and four bladed rotor. At a partcular rotor angle α, the blades had experenced forces due to pressure dfferences between the concave and convex surfaces, whch were resolved nto normal and tangental forces. As mentoned earler, the dfferences n pressure between the front and backsde of the blade result drag forces n normal and tangental drecton. Normal drag coeffcents are plotted aganst dfferent rotor angles n Fgure Angle of Rotaton Vs Total Normal Drag Coeffcent for Two,Three and Four Bladed Rotor Two Bladed Rotor Three Bladed Rotor Four Bladed Rotor Angle of Rotaton Two Bladed Rotor Three Bladed Rotor Four Bladed Rotor Angle of Rotaton Fgure 4. Varaton of the Total Tangental Drag Coeffcent wth the change of the Angle of Rotaton for Two, Three and Four Bladed Rotor The tangental drag co-effcent n case of two bladed rotor the value of C t fluctuates around.1 from to 4 then t starts decreasng up to 3. For both the three and four bladed rotor t starts decreasng at frst and reaches nto the hghest negatve co-effcent and then t starts ncreasng n a fluctuatng manner as the degrees ncreases.torque drag coeffcents are plotted aganst dfferent rotor angles n Fgure 5. Fgure 3. Varaton of the Total Normal Drag Coeffcent wth the change of the Angle of Rotaton for Two, Three and Four Bladed Rotor In case of two bladed rotor C n decreases from α= rotor angle up to α=1 n a fluctuatng manner 141

6 Torque Coeffcent Angle of Rotaton Vs Total Torque Coeffcent for Two,Three and Four Bladed Rotor Fgure 5.Varaton of the Total Torque Coeffcent wth the change of the angle of rotaton for Two, Three and Four Bladed Rotor The torque co-effcent s fluctuatng around the value of zero. It has the hghest postve coeffcent at α=8 and lowest negatve co-effcent at α=1 o. For the three bladed rotor the value of C q starts ncreasng from the α= and reaches the peak at α=4. After that t decreases abruptly up to α=6 and then t starts decreasng eventually n a fluctuatng manner and does not touch the zero lne. In case of four bladed rotor the value of C q ncreases from α= to α=6 and t has the hghest postve value at ths pont. Then t decreases drastcally up to α=1 and after that t starts ncreasng agan up to α=. After α= torque co-effcent decreases up to α=8 and at that pont t has the lowest negatve value. After α=8 t starts ncreasng agan. Overall the torque coeffcent shows a cyclcal pattern n decreasng manner for the four bladed rotor. 6. Concluson Two Bladed Rotor Three Bladed Rotor Four Bladed Rotor Angle of Rotaton The present study concerns the pressure dstrbutons over the two, three and four blades of the rotor on both the convex and concave surfaces at dfferent angles of rotaton. A pressure dfference s obtaned from the concave and convex surface, whch n turn gves drag coeffcents. The drag coeffcents may be resolved along the normal and tangental drectons of the rotor chord. The normal drag s responsble for the torque producton. The drag and hence the torque of each blade vares wth rotor angle. Further research s needed for studyng the effects on the performance of an S-shape savonus rotor by placng gude vanes around the rotor. The effects of the shape and the number of gude vanes may also be studed. Reference: [1] Lttler, R.D., Further theoretcal and Expermental Investgaton of the Savonus Rotor, B.E. Thess, Unversty of Queensland, Australa, [] Bowden, G. J. and McAleese, S.A., The propertes of solated and Coupled Savonus Rotors, Wnd Engneerng, vol. 8, no.4, UK, [3] Islam, M.Q., Islam, A.K.M.S., and Nur, M.A, Aerodynamc characterstcs of a statonary S-shaped Savonus Rotor, Dept of Mech. Engg., BUET. [4] [5] Islam, A.K.M.S., Islam, M.Q., Razzaque, M.M. and Ashraf. R.(1995), Statc Torque and Drag Characterstcs of an S-shaped Savonus rotor and Predcton of dynamc Characterstcs, Wnd Engneeerng, vol. 19, no 6. [6] Huda, M.D., Selm, M.A., Islam, A.K.M.S. and Islam, M.Q. (199), The performance of an S-shaped Savonus Rotor wth a Deflectng Plate, RERIC Int. Energy Journal, vol. 14 no. 1, pp [7] Lysen, E.H., Bos, H.G and Cordes, E.H (1978), Savonus Rotors for Water Pumpng, SWD publcaton, Amersfoort, The Netherlands. [8] Islam, M.Q. and Hossan, M.A., Savonus Rotor for Lftng Water n Bangladesh, Mech. Engg. Res. Bull., Vol.4 (1981), No