A PROTOTYPE DESIGN AND PERFORMANCE OF THE NEAL J. ROTH. B.A.Sc. (Mechanical Engineering), University of British Columbia,1982

Transcription

1 A PROTOTYPE DESIGN AND PERFORMANCE OF THE SAVONIUS ROTOR BASED IRRIGATION SYSTEM by NEAL J. ROTH B.A.Sc. (Mechanical Engineering), University of British Columbia,1982 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Mechanical Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA March, 1985 (c) Neal J. Roth, 1985

2 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make i t freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. NEAL J. ROTH Department of Mechanical Engineering The University of British Columbia 2975 Wesbrook Place Vancouver, Canada V6T 1W5 Date: March, 1985

3 i i ABSTRACT Important stages in the development of a wind energy operated irrigation system, which is simple in design and easy to maintain, are described from model tests in wind tunnels through to a prototype prepared for f i e l d tests. The attention is focussed on gross features of the protoype including the blade geometry and aspect ratio; mast, sleeve and bearing assemblies; braking system and a load matching concept. Described towards the end are the field test arrangements of the prototype and associated instrumentation. Even according to the most conservative estimate, the prototype tests suggest that the windmill should be able to deliver around 3000 l i t e r s of water per day (eight hours of wind) to a head of 5 m in a 24 km/h wind.

4 i i i TABLE OF CONTENTS Chapter Page 1 INTRODUCTION Preliminary Remarks A Brief Review of the Relevant Literature Scope of the Present Investigation 6 2 MODEL STUDIES Single-Stage Rotor Performance of the Single Stage Models Two-Stage Rotors The First Two-Stage Model and its Performance Test Procedure and Results for the Second Two-Stage Rotor 35 3 PROTOTYPE DESIGN System Assembly Braking System Transmission of Power and Load Matching System Monitoring Instrumentation and Controller Performance of Prototype 60

5 i v 4 CONCLUDING REMARKS 68 REFERENCES 70 APPENDIX I - MICROPROCESSOR BASED CONTROLLER 74 APPENDIX II - PREDICTION OF THE PROTOTYPE PERFORMANCE BASED ON WIND TUNNEL STUDIES 77 APPENDIX III - ASSEMBLY AND INSTALLATION OF THE PROTOTYPE. 79 APPENDIX IV - MATERIALS AND SUPPLIERS 87

6 V LIST OF FIGURES Figure Page 1-1 A schematic diagram of the wind energy operated irrigation system Blade configuration used to study gap-size, overlap and aspect ratio effects A schematic diagram showing model of the single-stage Savonius rotor A schematic diagram of the wind tunnel used in the single-stage model study Photograph showing a single-stage model during a typical test Prony brake arrangement for measurement of torque output Plots showing the effect of gap-size on power output at a given overlap and wind speed Effect of blade overlap on the rotor output for zero gap-size and a fixed wind speed 19

7 V 1 Figure Page 2-8 Variation of the maximum power coefficient with overlap showing the optimum setting of around 22% Plots showing effect of the aspect ratio (A = h/d) on the power coefficient of a single-stage Savonius rotor. Note, the aspect ratio of 0.77, which leads to a maximum Cp, was used in the prototype design A schematic diagram of the gear pump used in performance tests Performance chart for the single-stage Savonius model showing variation of flow rate with head and wind speed using a gear pump A schematic diagram of the large open c i r c u i t, blower type wind tunnel used for testing of the two-stage models Drawing showing the frame and f i r s t two-stage model in the large wind tunnel. Blade geometry is the same as that used in the single-stage study Effect of wind speed on power output and rotor rpm for the f i r s t two-stage Savonius model 29

8 v i i Figure Page 2-15 Variation of the power coefficient with tip speed ratio as affected by wind speed for the f i r s t two-stage model. The lack of collapse of the plots on a single curve is attributed to f r i c t i o n a l losses in the drive system Performance chart for the f i r s t two-stage rotor operating in conjunction with a gear pump A schematic diagram of the rotary pump used with the two-stage rotor Variation of flow rate with head and wind speed for the f i r s t two-stage Savonius rotor using a rotary pump 3A 2-19 Geometry of the blade profile used in the second two-stage model Support system for the second two-stage rotor Geometric details for the second two-stage model: 0 = 135 ; p/q = 1; A = Details of a dynamometer used with the second two-stage model ; AO

9 v i i i Figure Page 2-23 Variation of power with rpm at three different wind speeds for the second two-stage model Effect of wind speed on power coefficient vs. rpm plots for the second two-stage model Test arrangement for evaluating bearing losses Plots showing f r i c t i o n a l losses in the bearings of the second two-stage rotor Variation of the power coefficient with rpm accounting for f r i c t i o n a l losses A four-stage Savonius rotor based wind energy operated 2 irrigation system (projected area = 4.45 m ) during f i e l d tests indicating major subassemblies including the emergency brake, drive shaft, pumping unit and anchoring arrangement A schematic diagram showing the position of the windmill and meteorological tower atop the Mechanical Engineering Machine Shop building 50

10 ix Figure Page 3-3 Relative position of blades in four stages of the prototype Savonius rotor Details of the sleeve and bearing assembly Mast assembly Pneumatically actuated braking system showing air-cylinder, calipers and shoes Drive shaft and pumping arrangement showing load matching technique using three pumps ( PI, P2, P3 ) Pumps used and associated instrumentation for instantaneous and integrated flow measurements A schematic diagram showing the bottom half of the mast and details of the drive shaft assembly together with meteorological mast and instrumentation Typical plots showing the time histories of wind speed and flow rate for the prototype over three hours Plots showing the time histories of wind speed and flow rate for the prototype over 7.2 hours 64

11 X Figure Page 3-12 Variation of wind speed plotted against flow rate for the 7.2 hour history A detailed circuit diagram for the signal processor showing the comparator (CMOS 4081), invertor (CMOS 4049), binary counter (TTL 74193) and decoder (TTL 74154) 75 I I I - l Top of the mast showing the cap, guy-wires and a part of the rotor 79 III-2 Top bearing assembly: (a) outer bearing-housing; (b) inner bearing-holder and locknut; (c) assembly on the mast 81 III-3 Bottom bearing and brake drum: (a) details; (b) assembly on the mast 82 III-4 Power transmission to the drive shaft. Note also the brake plate assembly and guy-wire attachments III-5 Lifting sequence during installation of the rotor: (a), (b) crane attachment to the rotor at cap (c) lowering of the mast on the base plate; (d) rotor in position with guy-wires anchored 85

12 x i Figure Page III-6 Photograph showing relative position of the rotor and the meteorological mast atop the Mechanical Engineering Machine Shop 86

13 x i i ACKNOWLEDGEMENT Investigation was supported by the Natural Sciences and Engineering Research Council of Canada, Grant No. A Assistance of Mr. F. Knowles and Mr. J. Wiebe, Senior Engineering Technicians, in design and construction of the wind turbine is gratefully acknowledged. Assistance of Mr. Andrew Kwok is also gratefully acknowledged for the design and construction of the microprocessor which controls the windmill. A special thank you is extended to Dr. V.J. Modi for his time, guidance and help throughout the project.

14 x i i i LIST OF SYMBOLS a blade gap-size 2 A aspect ratio, h /S = h/d b blade overlap 0^ power coefficient, P/(l/2)pV^S D shaft diameter d blade diameter, 2r h blade height p,q,0 parameters defining blade geometry (Figure 2-19) P power output r blade rad ius S projected blade area, dh V wind speed p air density 1 rotor angular velocity X tip speed ratio, oor/v Ap change in pressure Q flow rate

15 1 1. INTRODUCTION 1.1 Preliminary Remarks Since the beginning of time, man has worshipped elements of nature, often out of fear, and has also strived to u t i l i z e them to advantage. In particular, light, water and wind represent the elements frequently exploited by the early man. It is uncertain just how long ago man f i r s t started to use wind to help power his daily work. However, there exist references to Chinese and Japanese windmills in use as far back as 2000 B.C.. Hammurabi, the emperor of Babylon, planned to use windmills to irrigate fields in 1700 B.C. and Hero of Alexandira, in 300 B.C., reported of a small windmill being used to operate an organ. Although i t is not known when the Persians started using windmills, however, historical records indicate their existence by 13A B.C. [1]. By the middle of the seventh century A.D. building of windmills was a well understood craft in Persia. In western Europe, records t e l l of windmills being used to grind grain and pump water since 1100 A.D.. By the fourteenth century, windmills were a common sight in England and throughout Europe. For nearly four centuries, until 1900, the Dutch used as many as 10,000 windmills. Towards the end of the nineteenth century, Denmark had a total of 3000 industrial windmills and 30,000 smaller types for homes and farms. Windmills f i r s t appeared in North America in the year 1620.

16 2 Their construction began in Virgina and extended up and down the Atlantic Coast. They were used mostly for grinding grain until the end of the eighteenth century. However, i t was not until the late 1800's that interest in windmills really took hold across America. Since 1880 more than six million windmills have been erected, primarily in the Midwest and Southwest regions. There are s t i l l approxiately 900 wind turbines in use today in Holland and around 100,000 in North America [2], More recently, with greater awareness of the limited supply of f o s s i l fuels, considerable attention has been focussed on the u t i l i z a t i o n of wind energy. The extensive amount of literature which has evolved has been reviewed by several authors including Hutter [3], Blackwell et a l. [4], Govind Raju and Narasimha [5], and others. A careful study of the literature reveals two interesting aspects [6-13]: (i) In general, investigations may be classified into two broad categories: (a) Laboratory scale model investigations, normally conducted at academic institutions by technically qualified personel, seldom evolve to a prototype stage for f i e l d tests and production. (b) Operational devices, assembled by environmental enthusiasts possessing limited technical background, which function as novelties at an uncertain efficiency. Being isolated devices, they also f a i l to reach

17 3 the production stage, ( i i ) Although several commercial organizations have focussed attention on development of large scale wind turbines in the megawatt range, considerable demand exists for systems with 1-10 kw capability to serve needs of rural communities. Design of such small to medium scale turbines has received relatively less attention, particularly with reference to applications aimed at irrigation of farms and lighting of small rural communities. Taking into account the foregoing aspects, the present project aims at the developement of a windmill, which is simple in design and easy to maintain, using industrial infrastructure readily available in rural areas of developing countries. The Savonius geometry promises to meet these objectives quite effectively. Although the Savonius rotor has a relatively low efficiency it presents several advantages over other configurations: (a) simple geometry and ease of construction; (b) in general seif-starting; (c) performance independent of the wind direction; (d) low starting wind speed; (e) easy to maintain; (f) relatively inexpensive in terms of material, constructional and maintenance costs. The classical Persian wheel for drawing water from wells, irrigation canals or rivers, which consists of a string of

18 4 buckets and is driven by a pair of bullocks, is used extensively in Africa, Asia and South American even today. Indonesia is one such country which is interested in adopting a relatively modern system for this purpose. During a recent v i s i t to U.B.C. (1978), o f f i c i a l s of the Indonesian National Institute of Aeronautics and Space suggested that a wind operated irrigation system would be ideal for their country. Indonesia, a nation of around 10,000 islands, with regular wind patterns, is well-placed to exploit this natural renewable energy source. 1.2 A Brief Review of Relevant Literature The amount of literature available in the general f i e l d of wind energy is indeed enormous as the subject has been evolving l i t e r a l l y for centuries. It is not intended here to review this vast body of information. As the configuration selected for study is basically Savonius in nature, attention is focussed on the literature pretaining to that specific geometry. Khan [14], who tested various Savonius rotor configurations, concluded that optimum overlap varied with the shape of the rotor. The presence of a gap between the rotor blades caused a decrease in the power coefficient while the peak efficiency did not vary with a change in wind speed. Ushiyama and his colleagues [11] also concluded that absence of a gap provided better results. They found a Bach type blade at an overlap of 30-50% produced a maximum power cofficient and that a guide vane was

19 5 quite effective in increasing output. Jones, L i t t l e r and Manser [16], in their study to identify the most promising rotor geometry, also indicated that the gap between blades was of a doubtful value and that vane shape alone governed the per formance. In an extensive study, Shanker [15] evaluated effects of the geometry and Reynolds number on the two and three bladed Savonius rotors. The results demonstrated an improvement in performance with an increase in the Reynolds number in the range tested. Furthermore, the two bladed rotors always out performed the three bladed geometry. Sivasegaram [ ] presented results of a series of investigations on the Savonius-type rotors. He concluded that although a rotor with several blades was not sensitive to blade geometry, i t did produce a slightly better performance than the conventional Savonius configuration. However, a two bladed rotor with improved sectional geometry exhibited a substantial increase in power coefficient. Sivasegaram [20] has also described an experimental investigation aimed at determining the optimum design of a concentrator-augmentation-system. Sabzevari [21] too has discussed effects of several types of concentrators and diffusers on performance of the Savonius geometry. Govinda Raju and Narasimha [22] reported a design based on the Savonius concept for pumping water. The system was labour intensive and used local materials and s k i l l s to the fullest. They found that a 'soft' design such as theirs delivered the amount of water that compared favourably with the performance of

20 6 more expensive commercial machines. Alexander [23] presented wind-tunnel blockage corrections based on tests with a series of Savonius rotors. He found that the corrections could be as large as 50% for the blockage ratio of 33%. The results were confirmed by testing a series of models in two wind-tunnels with different test-sectional areas. Alder [24] looked at the mean and periodic components of torque, drag and side force on the Savonius rotor. He found that the optimum efficiency of the rotor occurred at a tip-speed ratio of around Scope of the Present Investigation This thesis describes evolution of a wind energy operated irrigation system (Figure 1-1) from model tests in wind tunnels, for optimization of system parameters, to a prototype prepared for f i e l d tests. The emphasis throughout is on simplicity of design and ease of maintenance, with technology within the reach of developing countries. A detailed set of c r i t e r i a was established to serve as a guideline for a design of the system. The following features are considered desirable for an irrigation system appropriate for use in developing countries: (i) labour intensive design, based on local materials and s k i l l s ; ( i i ) performance independent of wind direction, with a

21 Figure 1-1 A schematic diagram of the wind energy operated irrigation system.

22 8 self-starting rotor that demands virtually no attention ; ( i i i ) simple design with construction, installation and maintenance consistent with the limited s k i l l s of a farming community; (iv) generation of approximately 250 watts of power equalling the performance of a pair of oxen; corresponding to the daily irrigation requirement of an average 6-8 acre farm in the amount of approxiately 3,000 l i t e r s per day; (v) stated output (250 watts) to be achieved at a minimum windspeed of 24 km/h and a head of 5 m. The project is divided into several stages: 2 (i) A scale model study of the Savonius rotor (0.12m projected area) in a wind tunnel with a systematic variation of the blade profile, gap size and overlap to arrive at an optimum configuration. ( i i ) Wind tunnel tests with a larger two-stage model of 2 the Savonius rotor ( 0.6m projected area) to assess influence of several other parameters not considered in stage ( i ), such as aspect ratio, end-plates and the self-starting characteristics. ( i i i ) Tests using a larger two-stage model of the Savonius 2 rotor ( 1.12m projected area ) with the optimum configuration as arrived at in stages (i) and ( i i ), together with several commercial pumps, (iv) Detailed design and construction of a prototype

23 9 consisting of materials readily available in an advanced industrial society such as Canada, (v) Field tests of the prototype to assess performance and structural integrity. Stages (iv) and (v) were added to the program as a result of interest shown by several Canadian agencies and farming groups in using the system for, besides irrigation, drainage, sewage disposal, servicing depleted oil wells and remote navigational beacons, etc. (vi) Technology transfer phase involving simplification of the design using materials readily available in Indonesia. (vii) Field tests in Indonesia to assess performance and establish maintenance procedures. This will also help in assessing the loss of efficiency due to design simplifications, (viii) Production of the two versions of the same design indicated above, one for use in the technologically advanced areas and the other in developing nations. The thesis describes progress made so far in the first five stages.

24 10 2. MODEL STUDIES Several families of single and two stage models were used in the experimental program to assess the effect of blade profile, gap-size, overlap and aspect ratio. The models were tested in several smooth flow conditions using two wind tunnels of different cross-sectional areas which were ideally suited for this class of studies. 2.1 Single-Stage Rotor Preliminary experiments with four different blade configurations suggested the one similar to that proposed by Khan [6] to be promisimg (Figure 2-1). Hence the blade gap-size, overlap and aspect ratio studies were confined to this geometry. Basically, two sets of models were used in the test-program. The f i r s t series of two-bladed models (Figure 2-2) with a maximum 2 blade diameter of 330 mm and a height of 301 mm (0.1 m projected area, A=0.91) were primarily designed to study the effect of gap-size and overlap. The blades, rolled into desired shapes, were constructed from 16 gauge aluminum sheet and supported by two plexiglas end-plates, 6.35 mm thick and 381 mm in diameter. Rigidity of the plates was augmented by reinforcing them with aluminum disks, 200 mm in diameter and 6.35 mm thick. A 25.A mm diameter, high precision straight shaft supported the blade

25 Figure 2-1 Blade configuration used to study gap- si ze, overlap and aspect ratio effects.

26 381 mm 12 End / Plate Blades 301 mm 203 mm 330 mm Figure 2-2 A schematic diagram showing model of the single-stage Savonius rotor.

27 13 assembly in a pair of *seif-aligning bearings. During testing, no vibration problems were encountered, even at rotational speeds as high as 1600 rpm. The second set of models systematically varied height/diameter ratio for a fixed projected area to assess the effect of aspect ratio. Essentially the same model construction and support procedures were used as described above. The models were tested in a low speed, low turbulence, return type wind tunnel with a test-section of 0.91 X 0.68 m. The air speed can be varied from 1-50 m/s with a turbulence level of less than 0.1%. A Betz micromanometer with an accuracy of 0.2 mm of water was used to measure pressure differential across the contraction section of 7:1 ratio. The spatial variation of velocity in the test-section is less than 0.25%. A line drawing of the tunnel used in this phase of the study is shown in Figure 2-3. A photograph (Figure 2-4) shows the model during a typical test. For each blade setting, the torque output was measured using a variation of the conventional Prony brake arrangement (Figure 2-5). A braided nylon string was passed around a pulley, 110 mm in diameter, mounted on the lower end of the rotor shaft. One end of the string carried a pan to which weights could be added while the other was fixed to the free end of a cantilever beam which had four strain gauges mounted near its root, two in compression and two in tension. The output signal from the strain gauges, forming a part of the Wheatstone bridge, was amplified using a Bridge Amplifier Meter (BAM) and measured by a digital voltmeter.

28 r Turning vanes 16.3 m H I j Figure 2-3 A schematic diagram of the wind tunnel used in the single-stage model study.

29

30 16 Pulley Nylon Thread Cantilever BAM Voltmeter Figure 2-5 Prony brake arrangement for measurement of torque output.

31 17 Sensitivity of the system was Nm. The rotor speed required to compute the power output was measured using a strobotac. 2.2 Performance of the Single-Stage Models Typical results for variation of the power output with rotor rpm for different blade separation is shown in Figure 2-6. The results clearly show that as separation 'a' is increased, the maximum power for a given overlap diminishes. The maximum power with zero separation between the blades was 50.7 W at 885 rpm. In general, an increase in gap-size tended to cause the maximum output to occur at a higher shaft speed for a given wind velocity. The corresponding effect of the blade overlap 'b' on the rotor output for a given gap-size and wind speed is presented in Figure 2-7. As can be expected, the overlap affected the output substantially, partly because of a change in the effective rotor diameter. More informative would be the variation of maximum power coefficient with percentage overlap as given in Figure 2-8, which shows an optimum to. exist at around b/d = 22% with a [C ] of p ma x Of considerable interest was the effect of the aspect ratio 'h/d' on the wind turbine performance as shown in Figure 2-9. The results suggested an optimum value of around The information proved to be useful in the prototype design discussed

32 RPM Figure 2-6 Plots showing the effect of gap-size on power output at a given overlap and wind speed.

33 Figure 2-7 Effect of blade overlap on the rotor output for zero gap-size and a fixed wind speed.

34 b/d, % Figure 2-8 Variation of the maximum power coefficient with overlap showing the optimum setting of around 22%.

35 Tip speed ratio, -77- Figure 2-9 Plots showing effect of the aspect ratio (A = h/d) on the power coefficient of a single-stage Savonius rotor. Note, the aspect ratio of 0.77, which leads to a maximum Cp, was used in the prototype design.

36 22 later. With the optimum blade setting established (essentially zero gap-size and 22% overlap), the wind turbine was connected to a gear pump (Figure 2-10) to obtain performance data in terms of flow rate and head as functions of wind speed. The head was varied by changing the height of a storage tank over a range of 1-5 m and the discharge rate determined by timing the accummulation of ten l i t e r s of water into a pail. The results are given in Figure As can be expected, because of the small size of the rotor, the wind speeds involved were relatively high. On the other hand, the attainable flow rates were rather attractive. The information served as a useful starting point in the design of the larger scale models. 2.3 Two-Stage Rotors In addition to providing the needed information concerning the optimum blade configuration, the single-stage model study emphasized the presence of a dead spot, i.e. when the blades are aligned with the wind, the rotor f a i l s to start without assistance. Thus, from self-starting consideration, i t was necessary to have a two-stage rotor, with blades in the individual stages oriented orthogonal to one another. Furthermore, results suggested that to generate even 100 Watts at a wind speed of 24 km/h would require a projected area of about m. Due to the uncertainties mentioned above, i t was decided

37 Figure 2-10 A schematic diagram of the gear pump used in performance tests.

38 V,km/hn J i i i i i FLOW RATE, L/min Figure 2-11 Performance chart for the single-stage Savonius model showing variation of flow rate with head and wind speed using a gear pump.

39 25 to conduct tests with larger two-stage models before proceeding with a prototype design. 2 2 Because of the size of the two-stage models (0.6 m and 1.12 m projected area) a larger wind tunnel with a test-section of 1.58 x 2.44 x 24.4 m was used in this part of the experimental program. Powered by a 93 kw (approximately 125 h.p.) motor driving a 2.44 m diameter, 16 bladed axial flow fan with pneumatic pitch control, the open circuit tunnel is capable of producing the maximum wind speed of 90 km/h. The tunnel was specifically designed to simulate thick ground boundary layers for industrial aerodynamics studies. However, this feature was not utilized in the present study as the models were placed close to the entrance of the test-section. The tunnel is schematically shown in Figure The First Two-Stage Model and i t s Performance The f i r s t of the two-stage models had a projected area of m. The blade diameter was 1.09 m with a height of the individual stage at 0.5 m. The blades were optimally designed with no gap and 22% overlap as suggested by the single-stage model studies. The blades were constructed from 22 gauge aluminum with plywood endplates, 1.17 m in diameter and 12.7 mm thick, and reinforced with 13 mm thick aluminum discs, 0.6 m in diameter. The whole assembly was supported by a 38 mm diameter ground shaft held in position by a pair of self-aligning bearings (Figure

40 Axivane Series 2000 Rotor, 2.44 m dia., 16 Cast aluminum blades, 125 h.p. electric motor, 175,000 cfm at 700 rpm, Fisher pneumatic variable pitch control 1 honeycomb and 4 screens in 4 x 4 m settling section Figure 2-12 A schematic diagram of the large open circuit, blower type wind tunnel, used for testing of the two-stage models. ro ON

41 ). Output of this larger model was measured using a conventional 12 volt automotive generator coupled to the main shaft through a system of pulleys, resulting in a step-up ratio of 1:8. The desired load was introduced by controlling the generator field current. Performance of the model was also tested in conjunction with several positive displacement pumps. The effect of wind velocity on the power output and rpm of the rotor is shown in Figure A substantial drop in the rotor speed is apparent, together with larger output even at a moderate wind velocity. Nondimensiona1 representation of the same results as given in Figure 2-15 brings to light an important aspect concerning the f r i c t i o n a l losses. In general, one would expect variation of the power coefficient with tip speed ratio to be independent of the wind speed. However, in the present case the power coefficient showed a marked increase with wind speed. This may be attributed to friction losses in the generator drive system which depend on the rotor speed and diminish with an increase in OJ in the range of interest here. Now, P = + P g and C P.g (1/2)PV 3 S, where: P total power produced by rot or ; power consumed by f r i c t ion ; P power produced by generator g

42 28 Frame n Top Bearing Wind Tunnel 8' by 5' test-section Blades P - S - Pulley Shaft E - End Plate Bottom Bearing 12 V Generator Windmill Test-Stand Pump '/A Figure 2-13 Drawing showing the frame and first two-stage model in the large wind tunnel. Blade geometry is the same as that used in the single-stage study.

43

44 TIP SPEED RATIO, rcu/v Figure 2-15 Variation of the power coefficient with tip speed ratio as affected by wind speed for the first two-stage model The lack of collapse of the plots on a single curve is attributed to frictional losses in the drive system. o

45 31 Thus, C = P[l - (P f /P)]/(l/2)pV 3 S 1 P» 6 = C [1 - (P f /P)], and C p = P/(l/2)pV 3 S. Recognizing that is almost independent of the wind speed (section 2.3.2) and that P^/P decreases with an increase in the wind speed, the trend indicated in Figure 2-15 is logical. Furthermore, continuous torque on the shaft due to two stages may also be responsible for improving the output. It was thought appropriate to evaluate potential of the wind turbine in terms of i t s pumping capability, particularly in the irrigation oriented application. Hence the tests were carried out in conjunction with a variety of pumps to establish their s u i t a b i l i t y for the intended purpose. Figure 2-16 shows the performance of the wind turbine in conjunction with a gear pump. In spite of large f r i c t i o n a l and fluid dynamic losses in the drive and piping systems, a flow rate of 250 1/h to a head of Am at a moderate speed of 16 km/h is indeed encouraging. One of the disadvantages of a gear pump was the priming requirement. This was eliminated by the rotary pump represented schematically in Figure Although the starting wind speed was a l i t t l e higher here (Figure 2-18, 22.A km/h), the flow rate was essentially independent of the head, at least over the range of in te re st. A comment concerning desirability of the rotor to run at near

46 1 r < Id X Wind Speed, km/h Performance with a gear pump J I I I I I I I L 0 I II FLOW RATE, L/min Figure 2-16 Performance chart for the first two-stage rotor operating in conjunction with a gear pump. LO N O

47 Figure 2-17 A schematic diagram of the rotary pump used with the two-stage rotor. 33

48 T r r < Wind Speed, km/h: Performance of a rotary pump J I I I I I I L 8 10 II FLOW RATE, L/min Figure 2-18 Variation of flow rate with head and wind speed for the first two-stage Savonius rotor using a rotary pump. U3

49 35 optimum speeds in different wind conditions would be appropriate. Furthermore, i t is also desirable to restrict the rotor speed to a safe value under stormy conditions to avoid overstressing. A scheme to implement the above requirements is under consideration. One possibility is to couple the wind turbine to a pump as well as a generator, which in turn charges a panel of storage batteries. Load on the generator may be controlled by governing the f i e l d current according to a strategy that involves excess speed and water level. The electrical storage system can be used to light a village a few hours each night. A simple control system can be designed to accomplish this inexpensively Test Procedure and Results for the Second Two-Stage Rotor The single stage rotor study clearly suggested that the relative magnitude of the straightline portion of the blade with respect to its radius of curvature was an important parameter in the design of an efficient blade geometry. Based on this observation a simple procedure for obtaining families of blade profiles was developed, as illustrated in Figure Note that through a systematic variation of the major variables p/q, 8, a, and b together with the blade aspect ratio h/d, an optimum combination can be established in an elaborate wind tunnel test program. The amount of information that can be generated through variation of even some of these parameters is considerable. In the construction of the second two-stage model, 6 was fixed at

50 gure 2-19 Geometry of the profile used in the second two-stage model 00 ON

51 The model had a projected area of 0.6 m, blockage of 16.4% and a p/q ratio of 1. Figure 2-20 shows a schematic diagram of the second two-stage rotor in its frame located in the large wind tunnel. The strain gauge based load measuring device mentioned earlier proved to be inadequate at larger outputs. It was therefore modified into a bigger dynamometer using two concentric cylinders, one of them free to undergo rotational displacement under the action of the torque transmitted through high viscosity o i l in the gap. The torque dependent displacement was measured through the cantilever mounted strain gauges as before. The details of the model and torquemeter are presented in Figure 2-21 and 2-22, respectively. Figure 2-23 illustrates how wind speed affects the power vs. rpm plots for the second two-stage rotor. Variation of the power coefficient with the tip speed ratio showed a slight increase at higher wind speeds (Figure 2-24). This may also be attributed to f r i c t i o n a l bearing losses in the dynamometer as explained before. It is anticipated that performance of such a drag device should be almost independent of the Reynolds number in the range 1.9 x x io 5. To help collapse the Cp vs. tip speed plots on a single curve, f r i c t i o n a l losses for the second two-stage rotor were obtained using the apparatus shown in Figure The bearings used to hold the rotor and dynamometer were fastened to a disk and mounted on a shaft which was spun by a variable speed motor. The disk was connected by a linkage to a flexible beam which had strain gauges mounted on i t to measure the force. This enabled

52 w/////////s//// s /////////...>/> / Bearing Two stage rotor Rotor support frame Tunnel Dynamometer support frame JZZZZZZZZZZZl zzz Flexible couplings Four-vane viscous flow dynamometer Strain gage torquemeter Balancing load a to minimize \ vibration Figure 2-20 Support system for the second two-stage rotor. oo

53 disk blade n 1.25 mm aluminum sheet used for disks and blades F' 480 mm Figure 2-21 Geometric details for the second two-stage model 9 = 135 ; p/q = 1; A = 0.75.

54 40 Stationary cylinder Oil drain Figure 2-22 Details of a dynamometer used with the second two-stage model.

55 RPM Figure 2-23 Variation of power with second two-stage model. rpm at three different wind speeds for the

56 0.25 Tip speed ratio, ^ Figure 2-24 Effect of wind speed on power coefficient vs. rpm plots for the second two-stage model. 4> to

57 Belt drive Variable speed drive b 3 Frame Bearings under test Strain gage transducer T Mr Figure 2-25 Test arrangement for evaluating bearing losses, -c-

58 44 the power consumed to be calculated from the torque at various speeds. The results are presented in Figure Cp vs. tip speed ratio is now plotted (Figure 2-27) for the second two-stage rotor with the frictional losses taken into account. It is apparent that the results do not quite collapse into a simple curve, suggesting a slight Reynold's Number dependency over the wind speeds considered.

59 NTN Bearing s ^+ Koyo Bearing / + + X RPM Figure 2-26 Plots showing frictional losses in the bearings of the second two-stage rotor.

60 0.3 O Q 0.2 O 2C + + o Cp Figure 2-27 Variation of the power frictional losses. coefficient with rpm accounting for

61 47 3. PROTOTYPE DESIGN With the optimum blade geometry, aspect ratio, rotor staging and pumping characteristics in hand, design of a f u l l scale wind turbine system was initiated. As stated previously, the i n i t i a l objective of this project was to design the system for use in the rural, farming communities in Indonesia. Accordingly, the guiding criterion which shaped the design were the following: simplicity of construction, operation and maintenance; u t i l i z a t i o n of locally available material; and use of the technology that is compatible with the rural environment. However, as the design progressed, interest expressed by farmers in Newfoundland, Quebec, and British Columbia suggested that there was a significant local demand for the windmill. It was therefore decided to design a prototype using relatively sophisticated materials readily available in an advanced industrial country such as Canada, bearing in mind i t s ultimate application in a rural society. Thus, with a technology transfer phase, essentially the same basic design would be able to serve the need of farmers in Canada as well as technologically developing nations. The prototype described here, which has been constructed and installed ( Figure 3-1, also APPENDIX III ), was specifically designed for f i e l d tests to correlate: (i) wind speed and direction information with the rotor rpm and pumping characteristics and;

62 Figure 3-1 A four-stage Savonius rotor based wind energy operated irrigation system (projected area = 4.45 m 3 ) during field tests indicating major subassemblies including the emergency brake, drive shaft, pumping unit and anchoring arrangement. 00

63 49 ( i i ) wind tunnel model test results with those of the prototype. These objectives, together with the f i e l d location of the prototype (atop the new Civil/Mechanical Engineering Building, Figure 3-2), has imposed several additional constraints and complications in design which would not be present in the actual application. 3.1 System Assembly Essentially the system consisted of a four-stage Savonius rotor, each stage 1.2 m in diameter and m high (Figure 3-3), built on an aluminum sleeve (Figure 3-4). The aluminum sleeve was fastened to aluminum bearing housings at each end. The housings held ball bearings which in turn were supported by the mast. Ball bearings were choosen because they produced the least rotational f r i c t i o n at a minimum cost. The axial load, due to the weight of the rotor, presented no problems for the ball bearings. The outside diameter of the mast and bearing holder required the bearings to have an inside diameter of at least 127 mm. Standard bearings of that size are able to support an axial force of at least 9000 Newtons. The mast, essentially a steel pipe 11.6 m t a l l, was built in three sections of 3m, 3m and 5.6m in lengths (Figure 3-5) with the outside diameter of 115 mm for the lower two sections reducing to 102 mm O.D. for the top section. The sections are

64 m 11.6m 8ml 7.3m 1 jwind anemometer signal..water flow meter signal - RPM signal wind speed and - direction signal - brake signal.^magnetic clutch signals DATA LOGGER SIGNAL PROCESSOR Figure 3-2 A schematic diagram showing the position of windmill and meteorological tower atop the Mechanical Engineering Machine Shop building.

65 1.5 m Figure 3-3 Relative position of blades in four stages of the prototype Savonius rotor.

66 52 B 125 mm *Nop bearing housing 178 mm O.D., 127 mm I.D. ball bearing 152 mm O.D., 146 mm I.D. aluminum pipe 178 mm O.D., 127 mm I.D. ball bearing 125 mm bottom bearing housing and pulley brake drum, 165 mm O.D. C A brake plate Figure 3-4 Details of the sleeve and bearing assembly.

67 m 102 mm O.D. steel pipe mm A brake plate 1 guy wire 115 mm O.D. steel pipe 11.6 m 102 mm O.D. steel pipe welded to one side of pipe mast 6.0 m 3.0 m mast base plate Figure 3-5 Mast assembly.

68 54 held together by flanges reinforced with short pieces, of steel pipe, 102 mm O.D., welded at the joints. Mounting the turbine on the sleeve significantly reduced the rotating inertia, thus lowering the starting speed. In the present case, weight of the four stage rotor was only 150 kg resulting mostly from the materials used in the construction of blades and endplates. Blades in the adjacent stages were staggered by 45 resulting in smoother operation. The mast was held in position by eight guy wires, four running from the mast cap and the other four from the region just below the brake-plate, connected to four roof anchors. The prototype was designed so that the rotor ( i.e. the light rotating section of the assembly ) slipped over the heavy, stationary anchored mast. In this way, even with a bearing or a housing failure, the rotor is prevented from breaking free from the mast. Also two sets of guy wires were provided to anchor the mast from both above and below the rotor which experiences the maximum aerodynamic forces. The guy wires were designed to hold the rotor in a 160 km/h wind. As a safety precaution, each cable was rated to hold the entire wind load by i t s self. As an added precaution, the guy wire anchors were placed through the concrete roof of the Machine Shop, and secured with metal plates inside of the build ing.

69 Braking System The Savonius rotor was provided with an emergency braking system to guard against a possible structural failure at high wind speeds ( > 50 km/h ). Essentially i t resembled the conventional automobile brake consisting of a drum with spring activated callipers carrying asbestos pads (Figure 3-6). The callipers were mounted on the brake plate (Section A-A) and the brake drum was bolted to the bottom bearing housing (Section C-C) as shown in Figure 3-4. The braking force was adjustable to a maximum of 2000 N through an appropriate pretensioning of the spring. The braking operation was governed pneumatically. Under a normal operating condition, a compressed air-line at 680 kpa pressurized the air-cylinder and the drum was released. At a predetermined c r i t i c a l wind speed, the air supply was switched off automatically through a microprocessor actuated controller (APPENDIX I ). The air cylinder was then allowed to bleed through a needle valve causing the piston (of the air-cylinder) to close the calipers and brake the drum. The needle valve was adjustable to control bleeding rate and hence the rate at which the brake was applied. The brake was constructed in such a way that i t did not depend upon an external power source for activation. With a break in the a i r - l i n e, an open valve or the air compressor switched off, the brake would be applied automatically by the spring as an emergency precaution.

70 air supply to operate air cylinder 690 KPa' air cylinder to disengage brake shoes from drum - 40 mm O.D. piston - 50 mm piston stroke 100 mm pins to fasten brake shoes to brake plate 50 mm h 50 mm + 50 mm l 25 mm Figure 3-6 Pneumatically actuated braking system showing air-cylind calipers and shoes.

71 Transmission of Power and Load Matching System The power was transmitted from the Savonius rotor to the pumps through a drive shaft (Figure 3-9). The power was relayed by three AA-Vee-Belts from the bottom bearing housing to the drive shaft which was designed in five sections. The f i r s t section permitted tensioning of the belts -while the second section had universal-joints at both ends to accommodate the motion of the f i r s t section. The next two sections also had universal-joints to correct for any misalignment in the shaft. The half inch bearings used to support the drive shaft are also self-aligning. At the bottom of the drive shaft a pulley transmitted the power to a system of pumps located at the mast base (Figure 3-7) to f a c i l i t a t e installation and servicing. It also avoids the problem of suction l i f t on the pumps. Two water lines run between the supply and receiving water tanks, the former located at the bottom and the latter supported on the mast at a desired height to provide a required head. Earlier wind tunnel investigations with model turbines operating with several different pumps showed the positive displacement type to be better suited to conditions of widely varying rpm encountered in practice. For a positive displacement pump, the flow rate varies linearly with rpm, however, the wind 3 turbine output is proportional to V. Thus for an efficient operation a load matching procedure is nesessary. In the present

72 58 Figure 3-7 Drive shaft and pumping arrangement showing load matching technique using three pumps (P1,P2,P3).

73 59 case this was accomplished by three pumps rated a 9, 22 and 44 1/min at 1750 rpm. Each of the pumps was provided with a magnetic clutch operated by a 12V D.C. supply, which was switched on through a microprocessor circuit at a predetermined wind speed. With the smallest pump permanently connected to the turbine, the starting wind speed was 14 km/h. In the actual application: (i) the mast and the guy wires w i l l be replaced by the locally constructed support frame; ( i i ) the pump w i l l be connected directly to the rotor sleeve; ( i i i ) the bottom tank w i l l be an irrigation ditch, well or river, and the top tank a separately constructed reservoir as shown in Figure Monitoring Instrumentation And Controller A review of the literature indicates that, in general, the correlation of results for a prototype wind turbines and i t s model is indeed quite scarce. This is particularly true for the Savonius configuration. The prototype with i t s instrumentation w i l l provide this v i t a l information. The f a c i l i t y permits correlation of rotor rpm, instantaneous and integrated discharge rates, wind speed and direction, and turbulence intensity. It w i l l be helpful in assessing the effectiveness of blade geometry, pump configurations, and several load matching techniques.

74 60 The system is provided with a digital rpm meter, two paddle wheel type flowmeters (Figure 3-8) with instantaneous and integrated discharge data outputs and a well proven G i l l anemometer, a l l connected to a multi-channel data-logger with variable data sampling capability (Figure 3-9). The wind speed, direction, flow rate, etc. information can be sampled at several preprogrammed rates and recorded on a magnetic tape for analysis. 3.5 Performance of Prototype The prototype and associated performance measuring instrumentation has been in operation since April A wide variety of weather conditions prevailed during the period which provided several opportunities to assess the turbines 1 s performance over a range of wind speeds. Typical three and seven hour records of wind speed and flow rate are presented in Figures 3-10 and Figure 3-11, respectively. The head was held fixed at 4m. The minimum sampling period for the data logger is one minute and i t was used for a l l the recordings. However, in most situations the wind speed fluctuations were observed to occur at higher frequencies. This in conjunction with the rotor's inertia tended to distort the performance records. It was concluded that a sampling period of 1-5 seconds should be used in future to obtain better correlations, particularly between the wind speed and the flow rate.

75 61 water up down mast 25 mm O.D. steel drivyhaft SC-Process Control Systems 'Maxigard* Model C3000 signal conditioner P1-JABSCO PUMP * P2-JABSCO PUMP P3-JABSCO PUMP * Campbell Scientific Inc CR21 Micrologger M1-315-PO Paddle wheel flosensor M2-315-TO Paddle wheel flosensor IT 1 -SIGNET SCIENTIFIC MK 375 R 0-8 GPM INDICATOR/TOTALIZER IT 2-SIGNET SCIENTIFIC MK 375 R 0-30 GPM INDICATOR/TOTALIZER Figure 3-8 Pumps used and associated instrumentation for instantaneous and integrated flow measurements.

76 bottom bearing housing and pulley 62. sized V-belts ^drive shaft pulley R. M. Young Company Gill propeller vane \ mast clampd= clamps W W Weston digital direction indicator Model 2462 clamp= Campbell Scientific Inc CR21 Micrologger clamp' Weston digital tach Model 2462 SC-/Process Control Systems "Maxigard* Model C3000 signal conditioner base plate Figure 3-9 A schematic diagram showing the bottom half of the mast and details of the drive shaft assembly together with meteorological mast and instrumentation.

77 i 1 1 > I I I Time, hours igure 3-10 Typical plots and flow rate showing for the the time histories of wind spee prototype over three hours.

78 flow rate for the prototype over 7.2 hours.

79 65 In Figure 3-10 the wind speed results were averaged over six minutes giving ten points over an hour. The averaging was done in three different ways: simple arithmetic mean ? 1/9 V=(V 1 + V 2 +V 3 +V 4 +V 5 +V 6 )/6; rms V=[(V^+V^+V^+V^+V^+V^)/6] u z ; and equivalent speed based on the average power input /? V=[(V^+V2+V3+V^+V^+Vg)/6]. The last approach appears to be a more rational way of presenting the wind data. As output of a positive displacement pump varies linearly with its rpm, arithmetic mean of the flow rate was used in plotting the results. Accounting for the rotor inertia, the flow rate responds to the wind speed fluctuations with a phase difference. With an increase in wind speed, the rotor inertia w i l l store angular momentum causing the flow rate to lag. This is apparent through a comparison of peaks marked a-a, b-b, c-c, etc. In Figure 3-11, the wind speed results based on the average power input, and the flow rate are plotted against time. As in the case of the f i r s t record the wind speed peaks do relate to the flow rate but with a phase shift due to the inertia effects of the rotor. No attempt at load-matching was made here and only the smallest pump of 9 1/min was in operation which, at this relatively low wind speed, was quite adequate. Accounting for the losses the results matched rather well with the wind tunnel test predictions (APPENDIX I I ). Such prototype results should prove to be extremely useful in assessing the full-scale losses and appropriate modifications in design to account for them.

80 66 The 72 points of windspeed and flow rate from Figure 3-11 are plotted against each other in Figure A simple least square f i t of the data gives the preformance relation for the windmill as: WIND SPEED = 1.6 x (FLOW RATE) Where the wind speed is in km/h and the flow rate is in 1/min. Note, the windmill starts operating at 14.1 km/h or approxiately 9 mph and w i l l pump 6 l i t e r s per minute to a head of 5m in a 24 km/h wind.

81 67 Flow Rate (liters/min) Figure 3-12 Variation of wind speed plotted against flow rate for the 7.2 hour history.

82 68 CONCLUDING REMARKS Based on the results obtained, the following general conclusions can be made: (i) The study suggests substantial effects of the blade separation, overlap and aspect ratio. The absence of gap between the blades resulted in a maximum output while influence of the overlap was particularly noticeble only for b/d > The optimum value of 'A' was found to be ( i i ) Prototype performance can be predicted using wind tunnel results with acceptable engineering accuracy provided: (a) there is a precise estimate of the losses in the drive system; and (b) faster sampling of the wind speed,thus reducing scatter in the data, is available. ( i i i ) The prototype has a starting wind speed of 14.1 km/h ( = 9 mph). 2 (iv) The prototype with a projected area of 4.45 m is anticipated to deliver at least 3,000 l i t r e s of water per day to a head of 5 m in a 24 km/h wind (24 km/h wind for 8 hours per day), (v) There is a considerable demand for wind turbines with a power output in the range of 5-10 kw, particularly in rural communities. The Savonius

83 69 rotor based irrigation system of moderate capability described here represents only a small step in evolution of such a design. The field tests with the prototype have provided performance information helpful in achieving this goal. However, the system is amenable to considerable improvement. In particular, efforts at: (i) improving the blade geometry; ( i i ) assessing the blockage and Reynolds number effects during wind tunnel tests; ( i i i ) finding alternate approaches for load matching, which are mechanically simple and economically attractive; (iv) simplifying the rotor and support structure design using the locally available materials of developing countries; (v) the development of analytical models for the prediction of prototype performance; should prove rewarding.

84 70 REFERENCES Ahmedi, G., "Some Preliminary Results on the Performance of a Small Vertical Axis Cylindrical Wind Turbine, " Wind Engineering, Vol. 2, No. 2, 1978, pp Savino, J.M., and Eldridge, F.R., "Wind Power," Astronautics and Aeronautics, Vol.13, No. 11, 1975, pp Hutter, V., "Wind Power Machines," NASA-TT-F-16195, February Blackwell, B.F., et a l., "Wind Energy-A Revitalized Pursuit," Sandia Laboratories, Report No , March Govind Raju, S.P., and Narasimha, R., "A Low-Cost Water Pumping Windmill Using a Sail Type Savonius Rotor," Department of Aeronautical Engineering, Indian Institute of Science, Report No. 79 FM 2, Bangalore January Eldridge, F.R., Wind Machines, National Science Foundation, U.S. Government Printing Office, Washington, D.C, U.S. Department of Energy/NASA, Wind Turbine Structural Dynamics, NASA Conference Publication 2034, NASA Lewis Research Center, 1977.

85 71 [8] Fantom, I.D. (Editor), Proceedings of the 2nd International Symposium on Wind Energy Systems, Vols. I, II, BHRA Publisher, Cranfield, U.K., [9] Proceedings of the AIAA/SERI Wind Energy Conference, AIAA Publisher, New York, N.Y., U.S., [10] Proceedings of the 16th Intersociety Energy Conference, Atlanta, Georgia, ASME Publisher, New York, N.Y., U.S.A., [11] Proceedings of the 17th Intersociety Energy Conversion Engineering Conference, Los Angeles, IEEE Publisher, New York, N.Y., U.S.A., [12] Proceedings of Energex '82, Regina, Editor: F.A. Curtis, Published by Solar Energy Society of Canada Inc., [13] Modi, V.J., Roth, N.J., and Pitalawala, A., "Blade Configuration and Performance of the Savonius Rotor with Application to Irrigation System in Indonesia," Proceeding of the 16th Intersociety Energy Conversion Engineering Conference, Atlanta, Georgia, U.S.A., August 1981, ASME Publisher, Vol. 2, pp [14] Khan, H.M., "Model and Prototype Performance Characteristics of Savonius Rotor Windmill," Wind Engg., Vol. 2, No. 2, 1978, pp

86 72 [15] Shankar, P.N., "The Effects of Geometry and Reynolds Number on Savonius Type Rotors," Report AE-TM3-76, National Aeronautical Laboratory, Bangalore, [16] Jones, C.N., Litter, R.D., and Manser, B.L., "The Savonius Rotor - Performance and Flow," Preceedings of the First BWEA Wind Energy Workshop, Multi-Science Publishing Co. Ltd., The Old M i l l, Dorset Place, London E15 1DJ, U.K., April [17] Sivasegaram, S., "An Experimental Investigation of a Class of Resistance-Type, Direction-Independent Wind Turbine," Energy, Vol. 3, 1978, pp [18] Sivasegaram, S., "Design Parameters Affecting the Performance of Resistance-Type, Vertical-Axis Wind Rotors; An Experimental Investigation," Wind Engineering, Vol. 1, No. 3, 1977, pp [19] Sivasegaram, S., "Secondary Parameters Affecting the Performance of Resistance-Type Vertical-Axis Wind Rotor," Wind Engineering, Vol. 2, No. 1, 1978, pp [20] Sivasegaram, S., "Concentration Augmentation of Power in a Savonius-Type Wind Rotor," Wind Engineering, Vol. 3, No. 1, 1979, pp

87 73 [21] Sabzevari, A., "Power Augmentation In A Ducted Savonius Rotor," Proceedings of the 2nd International Symposium on Wind Energy System, Amsterdam, Oct. 1978, Paper F3, pp [22] Govinda Raju, S.P., and Narasimha, R., "Windmills for Rural Use," Department of Aeronautical Engineering, Indian Institute of Science, Report No. 77 FM 14, Bangalore, [23] Alexander, A. J., " Wind Tunnel Corrections For Savonius Rotor," Proceedings of the 2nd International Symposium on Wind Energy Systems, Amsterdam, Oct. 1978, Paper E6, pp [24] Alder, G.M., "The Aerodynamic Performance of The Savonius Rotor," Proceedings of the 2nd International Symposium on Wind Energy System, Amsterdam, Oct. 1978, Paper F3, pp [25] Modi, V.J., Fernando, M.S.U.K., and Roth, N.J., " An Approach to Wind Energy Operated Irrigation System, " Proceedings of Energex, the Global Energy Forum '84, Regina, Sask., Canada, May 1984, Editor: F. Curtis, pp

88 74 APPENDIX I - MICROPROCESSOR BASED CONTROLLER It was considered essential to incorporate a microprocessor based controller (Figure 1-1) to: (i) automatically apply the brake at a desired preset wind speed to assure safety of the structure; ( i i ) permit load matching by actuating a magnetic clutch thus connecting or disconnecting a pump according to the wind speed; ( i i i ) delay the activating signal by a desired amount to avoid sensitivity to transient fluctuations in the wind speed. The processor is able to achieve the above requirements through a comparator, clock pulse generator and a delay time counter. A comparator generates a constant signal of about 3-5 V when and input signal exceeds a reference value. In the present case an ' AND GATE' is used to act as a comparator. A CMOS 4081 chip provided a triggering voltage level of 2.5 V. Output of the G i l l anemometer was adjusted accordingly using a potentiometer to trigger a signal for control. For a signal delay system to operate satisfactorily, a clock pulse generator is required to provide a time reference. This can be achieved in a number of ways. However, for simplicity, in the present case i t was accomplished through the use of two inverters giving a pulse rate of (1.4RC)" 1. With R = 6.5Mfi and C = 0.22 pf,

89 +5v B B H l I M i l l i i i i 741 i i S3N I I I! * n» «L N -» - O s - S» 741S4N IK brake activation - o»» o» -» br «. k f ^clutch testj test IK 1 clutch activation.j ii 6V brake calibration* 1M i : IK IK *-W- 1K IM clutch calibration to brake GRD anemometer to clutch Figure 1-1 A detailed circuit diagram for the signal processor showing the comparator (CMOS 4081). invertor (CMOS 4049). binarv counter (TLL 74193) and decoder (TTL 74154).

90 76 a pulse is generated every 2 seconds. The chip used for the purpose was CMOS The time delay system consists of a 4-bit binary counter (TLL 74154), and a 'NAND GATE' (CMDS 4011). The function of the binary counter is to receive a signal from the clock pulse generator and convert the count in the binary form for the decoder. The decoder is essentially a decade counter which takes a binary number and converts i t into the decimal form. The 'NAND GATE' is used to terminate counting, by the binary counter, at a preset delay time, in turn controlled by switches connected to the output of the decoder. In a normal operation a l l switches w i l l be open except the one giving the desired time delay. Thus when the binary counter is triggered by the comparator, i t initiates the count of the clock pulses which serve as an input to the decoder. Now each 4-bit address from the binary counter drives one consecutive output of the decoder 'low' keeping other outputs 'high'. Hence, at this stage, the output of the decoder would go 'low' consecutively from N = 1. For example, i f switch number 8 is closed, then the input to 'NAND GATE' is high because the N = 8 output is s t i l l high when the counting i n i t i a t e s. It w i l l take 8 clock pulses before the N =8 output goes low together with the input to the 'NAND GATE'. As the low N = 8 output of the decoder also goes through an inverter, the output from the latter would be high. The signal can then be used to activate external devices like brakes and clutches.

91 77 APPENDIX II - PREDICTION OF THE PROTOTYPE PERFORMANCE BASED ON WIND TUNNEL STUDIES Consider the two-stage model with a projected are of 1.2m as used in the wind tunnel test shown in Figure 2-13 (page 28). The corresponding output in conjunction with a positive displacement rotary pump is presented in Figure 2-18 (page 34). The plots show a flow rate of 6.3 1/min (1.05 x IO - 4 m/s) to a head of 5m ( N/m ) at a wind speed of 22.4 km/h. 2 Power = Q x Ap = (1.05 x 10 A )( x 10 4 ), = 5.13 W. Now C p = Power/(l/2)pV 3 S, where: p = kg/m ; V = 22.4 km/h = S = 1.2 m. m/s; Therefore C = P The rotor presented a blockage of 27.4%. Correcting for the blockage as suggested in reference [25], the effective coefficient for the model i s, C = (0.476)., _ x (0.029) p correction factor =

92 78 Turning to the prototype with a projected area of 4.45 m, Power = (0.0138)(l/2)pV 3 S. Taking V 20 km/h = 5.55 m/s, Power = 6.A3 W. There fore Q = Power/Ap = 6.A3/A89A5 = 1.31 x 10 m/s, = 0.47 m 3 /s. With the wind available for an average of 8 hours per day, this w i l l amount to, Q = /day. From Figure 3-12 (page 67) for 20 km/h wind, Q = 3.8 1/min, i.e /day (8 hours). This is almost half of the predicted value. However, following factors must be kept in mind: (i) steep gradient of the head vs. flow rate plot in Figure 2-18; ( i i ) uncertainty of the blockage corrections; ( i i i ) additional losses (compared to the model) due to the seven bearings used to support the drive shaft and the four V-belts used to transmit the power; (iv) limitations of the data-logger in terms of sampling rate (maximum one sample per minute) resulting in highly scattered wind data; (v) linear, least square f i t of the scattered data. Considering these points, the prediction is indeed encourging.

93 79 APPENDIX III - ASSEMBLY AND INSTALLATION OF THE PROTOTYPE The system involved a large number of parts and subassemblies each requiring careful design, fabrication and integration. The entire welding and machining operations were done in the department's machine shop with major (60 %) particapation of the author in collaboration with two machine shop technicians. The rotor blades were rolled into the desired shape by the author using a hand-roller from the C i v i l Engineering Department. The rotor blades and end-plates were secured with pop-rivets. Lining of the brake shoes was done commerica11y.

94 80 The mast was assembled by holding i t one meter away, parallel to the ground, on two end-stands. The top end of the mast was provided with a mast cap to f a c i l i t a t e l i f t i n g and anchoring of the system using guy-wires ( Figure I I I - l ). The mast supports the rotor at its ends through two similiar bearing assemblies. The bottom bearing assembly was installed on the mast f i r s t. Next, the rotor was slipped over the mast followed by the top bearing assembly. Figure III-2a shows the top outer bearing housing while Figure III-2b presents the top mast bearing-holder together with the locknut. The entire bearing assembly appears as indicated in Figure III-2c. Details of the bottom bearing assembly showing the seal and the brake drum are presented in Figure III-3a. The teflon seal prevents draining of the grease between the races due to gravity. Figure III-3b shows the assembled brake drum together with bearing. Power from the rotor is transmitted to the pumps located at the bottom of the mast through V-belts and a drive shaft (Figure III-A). Note the brake plate assembly in blue supports the brake shoes and provides for anchoring of guy-wires as well. Installation of the rotor, being on the roof of the Mechanical Engineering Machine shop, proved to be somewhat challenging and involved careful preplanning. Arrangements were made with the Physical Plant to embed eight anchors and a base plate (to support the mast) in existing concrete beams. One-half inch diameter guy-wires and the mast were designed to withstand 160 km/h winds. This was primarily a precaution against damaging

95 81 Figure III-2 Top bearing assembly: (a) outer bearinghousing; (b) inner bearing-holder and locknut; (c) assembly on the mast.

96 Figure III-3 Bottom bearing and brake drum: (b) assembly on the mast. (a) det aiis ;

97 83 Figure III-4 Power transmission to the drive shaft. Note also the brake plate assembly and guy-wire attachments. the building. On the other hand, to simulate field conditions more realistically, the blades where designed for a drag load of only 80 km/h to simplify construction and reduce cost. The erection was executed by Johnston Contract and Heavy Haul Service using a 65 ton crane. The rotor, with one end of the guy-wires attached to the mast, was lifted at the mast cap (Figures III-5a and III-5b) and lowered on the base plate (Figure III-5c). The guy-wires were now attacted to the roof anchors and adjusted to achieve vertical alignment of the rotor (Figure III-5d). With the rotor in place the pumps, tanks, flow and rpm

98 84 measuring transducers were installed. The airline for actuation of the brake and wires conveying signals, protected by a PVC conduit, were connected to the display instrumentation located downstairs within the building (ME 372 laboratory). A meteorological mast with a Gill anemometer was erected on the roof about 20m south of the rotor. The rotor, meteorological mast and the surrounding terrain are shown in Figure III-6. (a) (b) Figure III-5 Lifting sequence during installation of the rotor: (a), (b) crane attachment to the rotor at cap.

99 85 Figure III-5 Lifting sequence during installation of the rotor: (c) lowering of the mast on the base plate; (d) rotor in position with guy-wires anchored.

100 Figure III-6 Photograph showing relative position of the rotor and the meteorological mast atop the Mechanical Engineering Machine Shop. 00 cr-