The development of a segmented variable pitch small horizontal axis wind turbine with active pitch control

(1)

_____________________________________________________________

COPYRIGHT STATEMENT

The copy of this thesis has been supplied on condition that it is understood that the copyright rests with the author and that no information from it may be published without the consent of the author.

(2)

_____________________________________________________________ THE DEVELOPMENT OF A SEGMENTED VARIABLE PITCH SMALL HORIZONTAL AXIS WIND TURBINE WITH ACTIVE PITCH CONTROL

By

SEAN POOLE

A thesis submitted in fulfilment of the full requirements for the degree of MASTER DEGREE: ENGINEERING: MECHATRONIC

Faculty of Engineering, the Built Environment and IT, Nelson Mandela Metropolitan University

December 2013

Supervisor: Prof Russell Phillips Co-supervisor: Prof Theo van Niekerk

(3)

_____________________________________________________________

ABSTRACT

Small scale wind turbines operating in an urban environment produce dismal amounts of power when compared to their expected output [1-4]. This is largely due to the gusty wind conditions found in an urban environment, coupled with the fact that the wind turbines are not designed for these conditions. A new concept of a Segmented Variable Pitch (SVP) wind turbine has been proposed, which has a strong possibility to perform well in gusty and variable wind conditions.

This dissertation explains the concept of a SVP wind turbine in more detail and shows analytical and experimental results relating to this concept. Also, the potential benefits of the proposed concept are mentioned.

The results from this dissertation show that this concept has potential with promising results on possible turbine blade aerofoil configurations. Scaled model tests were completed and although further design optimisation is required, the tests showed good potential for the SVP concept. Lastly a proof-of-concept full scale model was manufactured and tested to prove scalability to full size from concept models. Along with the proof-of-concept full scale model, a wireless control system (to control the blade segments) was developed and tested.

(4)

_____________________________________________________________

LIST OF FIGURES

Figure 1: Effect of Building on Wind Leeward of Structure [4] ... 2

Figure 2: Operating TSR of a 3m HAWT in Gusty Wind Conditions ... 3

Figure 3: Concept Rendering of SVP Wind Turbine ... 7

Figure 4: Concept Diagram ... 8

Figure 5: CAD Model of Conceptual Test Prototype ... 8

Figure 6: Flying Wing Concept ... 9

Figure 7: Tailed and Canard Test Models ... 10

Figure 8: Wind Tunnel Setup ... 11

Figure 9: Force Balance Diagram ... 12

Figure 10: Gauge for Setting/Reading the Angle of Attack... 12

Figure 11: Aerofoil Results (Open-Loop Wind Tunnel) ... 13

Figure 12: NACA 4412 Aerofoil ... 13

Figure 13: NACA 0012 Aerofoil Configuration... 14

Figure 14: AOF139 Aerofoil ... 14

Figure 15: Diagram of NACA0012 Aerofoil Configuration ... 16

Figure 16: Stability Graph ... 17

Figure 17: AOF139 Coefficient of Moments ... 18

Figure 18: Adjusted Coefficient of Moment (Pivot Point Foreword of Aerodynamic Centre) ... 18

Figure 19: Lift to Drag Ratios at the Predicted Stable Points ... 19

Figure 20: Test NACA 4412 Aerofoil Configuration in Wind Tunnel ... 19

Figure 21: Test NACA0012 Aerofoil Configuration in Wind Tunnel ... 20

Figure 22: Test AOF139 Aerofoil in Wind Tunnel ... 21

Figure 23: AOF139 Shaft Locating Templates ... 21

Figure 24: CAD Model of Tested NACA4412 Aerofoil ... 22

Figure 25: Tested NACA 4412 Aerodynamic Coefficients at Re=50 000 ... 22

Figure 26: Tested NACA 4412 Lift to drag Ratios at Re=50 000 ... 23

Figure 27: CAD Model of Test NACA 0012 Aerofoil Configuration ... 23

Figure 28: Tested NACA 0012 Configuration Aerodynamic Coefficients at Re=50 000 ... 24 Figure 29: Tested NACA 0012 Configuration Lift/Drag Ratios at Re=50 000 . 24

(7)

_____________________________________________________________ Figure 31: Tested NACA 0012 Configuration Aerodynamic Coefficients Vs.

Tail Angle at Re=50 000 ... 25

Figure 32: Tested NACA 0012 Configuration Lift/Drag Ratios Vs. Tail Angle at Re=50 000 ... 26

Figure 33: CAD Model of Tested AOF139 Aerofoil ... 26

Figure 34: Tested AOF139 Aerodynamic Coefficients (Including 4 Unfixed Test Results) at Re=50 000 ... 27

Figure 35: Tested AOF139 Stable and Unstable Aerodynamic Coefficients Vs. Lapsed Time at Re=50 000 ... 28

Figure 36: Tested AOF139 Lift/Drag Ratios at Re=50 000 ... 28

Figure 37: Tested AOF139 Angle of Attack vs. Pivot Point Position at Re=50 000 ... 29

Figure 38: Scaled Model Test Setup ... 31

Figure 39: Arduino Uno Connected to Interfacing Board ... 31

Figure 40: MOSFET Bank [19] ... 32

Figure 41: Setup ... 32

Figure 42: Inductive Proximity Sensor Setup [20] ... 33

Figure 43: Generator Equivalent Circuit ... 34

Figure 44: Mechanical Torque Measurement Setup ... 35

Figure 45: Measured Data and Model ... 36

Figure 46: Constant K as a function of Duty Cycle ... 37

Figure 47: Hand Tachometer ... 37

Figure 48: Arduino PWM Signal ... 38

Figure 49: Aerofoil Diagram ... 41

Figure 50: CAD Model of Fixed Two Blade Constant Chord Turbine ... 43

Figure 51: Printed Turbine Blade Components ... 43

Figure 52: CAD Model of Fixed Two Blade Variable Chord Turbine ... 44

Figure 53: Fixed Blade Turbine Power Curve at a Wind Speed of 4.5m/s ... 45

Figure 54: Lift/Drag Ratio Comparisons [23]. ... 46

Figure 55: CAD Model of Two Blade Segmented Variable Pitch Turbine ... 47

Figure 56: Variable Pitch Results (Wind velocity = 4.5m/s) ... 48

(8)

_____________________________________________________________ Figure 59: Three Blade SVP Turbine "Fixed" at Various Positions (Wind

velocity = 4.5m/s)... 50

Figure 60: Trailer Test Rig ... 52

Figure 61: Segmented Blades Attached to a Ginlong GL-PMG-1500 Generator ... 53

Figure 62: Tow Test Trailer Setup ... 53

Figure 63: Servo Motor ... 54

Figure 64: Servo Driver Circuit Diagram ... 55

Figure 65: Servo Driver ... 55

Figure 66: Servo Driver Minimum Output (1ms Pulse) ... 56

Figure 67: Servo Driver Maximum Output (2ms Pulse) ... 56

Figure 68: 433Mhz RF Link Kit with Encoder and Decoder [30, 31] ... 57

Figure 69: 433Mhz RF Link Kit [32, 33] ... 57

Figure 70: Transmission Lag (433Mhz RF Link with Encoder/Decoder) ... 58

Figure 71: Wireless Control Diagram (On-board Servo Signal Generation) ... 58

Figure 72: Solid State Relay Controlled Variable Resistor Bank – R(A) ... 59

Figure 73: Wireless Control Diagram (Transmitted Servo Signal) ... 60

Figure 74: Manually Switched Variable Resistance Bank ... 61

Figure 75: Transmitted and Received Data... 62

Figure 76: Static Signal Output from Receiver ... 62

Figure 77: Realised Receiver Setup ... 63

Figure 78: Realised Transmitter Setup ... 63

(9)

_____________________________________________________________

LIST OF TABLES

Table 1: Properties of Constant Chord Fixed Blade Turbine ... 42 Table 2: Properties of Variable Chord Fixed Blade Turbine ... 44 Table 3: Properties of SVP Blade Turbine ... 47 Table 4: Calculated Resistance Values for Solid State Relay Switched Resistance Bank ... 60 Table 5: Calculated Resistance Values for Manually Switched Resistance Bank ... 61 Table 6: Properties of Full Scale Turbine ... 64

(10)

_____________________________________________________________

GLOSSARY OF TERMS

A

Aerodynamics – refers to the physics of motion of air especially when it interacts with a moving object.

Aerodynamic Centre – the point on an aerofoil where the pitching moment coefficient does not change with varying angle of attack.

Aerofoil – a structure with curved surfaces designed to give the most favourable ratio of lift to drag (Also referred to as an airfoil).

AC (Alternating Current) – electricity in which the electric current periodically switches direction, generally in the form of a sine wave.

Arduino (Uno) – a microcontroller board based on the ATmega328 chip. ASK (Amplitude Shift Keying) - a modulation method in which the amplitude of the carrier signal is varied to represent binary 0 or 1.

Asymmetrical – refers to the absence of symmetry. B

Blade Element Momentum Theory (BEM Theory) – is a mathematical model applied to the design of wind turbine rotors.

C

CAD (Computer Aided Design) - the use of computer technology for the process of creating computer simulated 2D or 3D models.

Camber – the asymmetry between the top and the bottom surfaces of an aerofoil.

Capacity Factor – the ratio of average power produced relative to maximum rated power.

CD (Drag Coefficient) – a dimensionless coefficient that relates the drag

generated by an aerodynamic body to the density of the fluid around the body, its velocity and an associated reference area.

Chord – straight line measurement of the length of an aerofoil cross section from the tip of the leading edge to the tip of the trailing edge.

CL (Lift Coefficient) – a dimensionless coefficient that relates the lift

(11)

_____________________________________________________________ CM (Moment Coefficient) – a dimensionless coefficient that relates the

moment generated by an aerodynamic body to the density of the fluid around the body, its velocity and an associated reference area.

Computational Fluid Dynamics (CFD) - is one of the branches of fluid mechanics that uses numerical methods and algorithms to solve and analyse problems that involve fluid flows.

D

Duty Cycle – the ratio of the duration of a square wave signal pulse relative to the period.

F

FEA (Finite Elements Analysis) – a branch of structural analysis that uses numerical methods and algorithms to solve and analyse problems.

Fenestron – a ducted fan tail rotor of a helicopter.

Free Stream Velocity – the velocity of the air (wind) far upstream of a wind turbine before it is disturbed.

G

Generator – a device that converts mechanical energy into electrical energy. H

Horizontal Axis Wind Turbine (HAWT) – a wind turbine with the rotor rotating about a horizontal axis.

Hub – the central portion of the rotor to which the blades are attached. The hub also serves the purpose of mounting the rotor to the generator shaft.

M

MOSFET (Metal-Oxide Semiconductor Field-Effect Transistor) – a field-effect transistor consisting of a thin layer of silicon oxide between the gate and the channel.

MPPT (Maximum Power Point Tracker) – a device (in the context of wind energy) used to get the maximum possible power from a wind turbine by varying the load applied to the generator to achieve optimum turbine speed.

(12)

_____________________________________________________________ N

NMMU (Nelson Mandela Metropolitan University) – a university based in Port Elizabeth, South Africa.

P

PWM (Pulse Width Modulation) – a technique of modulating the duty cycle of a set frequency square wave pulse.

R

Renewable Energy – an energy source which comes from natural sources such as the sun, wind, water, tides and geothermal heat.

Reynolds Number (Re) – a dimensionless number that gives a measure of the ratio of inertial forces to viscous forces in a flowing fluid.

RF (Radio Frequency) – electromagnetic waves between 3kHz and 300GHz. Rotor Solidity – ratio of the blade surface area to swept area.

Rudder – a device for governing an objects direction. S

Span – distance from blade root to tip or similarly the length of an aerofoil (perpendicular to aerofoil profile).

T

TSR (Tip Speed Ratio) – the ratio of the speed of the tip of the wind turbine blade relative to the speed of the free stream velocity of the wind.

Turbine – refer to wind turbine.

W

Wind Energy – refers to the kinetic energy available in the wind.

Wind Turbine – a rotating device that converts kinetic energy from the wind into mechanical energy.

X

XFOIL – an analysis and design software for low Reynolds number aerofoil characterization.

(13)

_____________________________________________________________ Y

(14)

_____________________________________________________________ NOMENCLATURE

A – Area of an Aerofoil (Span x Chord) [m2] a – Axial Induction Factor [1]

a‟ – Tangential Induction Factor [1] C – Chord Length [m] or Capacitance [F] CD – Coefficient of Drag [1] CL – Coefficient of Lift [1] CM – Coefficient of Moment [1] CP – Coefficient of Power [1] D – Drag [N] DC – Duty Cycle [%] F – Force [N] f – Frequency [Hz]

I – Direct or RMS Current [A] L – Lift [N] or Length [m] M – Moment [N·m] N – Number of Blades [1] P - Power [W] R – Radius [m] or Resistance [ ] Re – Reynolds Number [1] T – Torque [N·m] Tr – Local Thrust [N] Qr – Local Torque [N·m] V – Velocity [m/s] or Voltage [V] α – Angle of Attack [°]

αr – Local Angle of Attack [°]

β – Air Flow Angle [°]

βr – Local Air Flow Angle [°]

γr – Local Blade Pitch [°]

λ – Tip Speed Ratio (TSR) [1] λr – Local Speed Ratio [1]

(15)

_____________________________________________________________ Ø – Diameter [m]

(16)

_____________________________________________________________

ACKNOWLEDGEMENTS

I would hereby like to acknowledge the following people who were instrumental in the completing of this research dissertation:

My supervisor, Prof Russell Phillips, for his guidance and support, and financial backing.

My co-supervisor, Prof Theo van Niekerk, for his advice and guidance. My wife, Jade, for her support, both financial and moral.

The Renewable Energy Team at NMMU, especially Leeroyd Sophola who helped me with much of the programming.

Gideon Gouws, for his advice and for his timely designing and constructing the wind tunnel at NMMU of which much of my research would not have been possible.

To my Lord and saviour, for providing me the opportunity of coming back to university and completing my studies.

(17)

_____________________________________________________________

AUTHOR‟S DECLARATION

I hereby declare that the work contained in this thesis is my own original work and that I have not previously submitted it, in its entirety or in part, at any university for a degree.

____________________ SEAN POOLE

(18)

Chapter 1 Project Scope _____________________________________________________________

CHAPTER 1: PROJECT SCOPE

CHAPTER 1 PROJECT SCOPE

1.1 BACKGROUND

History: The use of wind turbines is by no means a new concept; it has been around for at least three thousand years [5]. In Scotland, windmills were in use since the early 15th century to mill grain and pump sea water [5]. Then in the 19th century there was a boom in wind pumps in the USA which saw over six million wind pumps installed between 1850 and 1970 [5]. 1887 saw the first electricity generated using wind by means of a wind turbine driven electrical generator [5]. It took a further 54 years for the first wind powered electrical generator of megawatt-scale to be developed (Putnam-Smith wind turbine in 1941 [5, 6]). There was not to be another electrical wind turbine of this size for another 40 years simply because this form of power generation was too expensive [5]. The oil crisis in 1973 reignited interest in wind power with growth ever since, and as of January 2001, a global capacity of about 16.5 GW was installed, with about half of that in Europe [5].

Small scale wind power: As the global economic trend for large wind turbines increases, so is the trend for small privately owned wind turbines. As individuals become more 'green' conscience they are searching for a viable and sustainable alternate energy supply. Since small wind turbine costs are decreasing, and the cost of supplied electricity increasing, this has allowed more people to consider investing in small scale wind turbines for private electricity generation. But the performances of small wind turbines in urban areas has proved dismal compared to the claim power production by the manufacturer [1-4]. This is largely due to the wind conditions found in urban areas, which are gusty (in both speed and direction) and also not as strong. The wind in urban areas is buffeted by buildings, structures, and trees, as shown in Figure 1.

(19)

Figure 1: Effect of Building on Wind Leeward of Structure [4]

The Warwick Wind Trials are a testament to this fact, where 26 small wind turbines situated in predominantly urban areas were performance tested. These independently owned turbines were tested over a period of a year (2007-2008). The average capacity factor for these turbines was found to be 0.85% (4.15% excluding down time) with some turbines' electronics using more power than what was generated [1].

Another study done in the UK by Energy Saving Trust on small scale turbines operating in urban environments suggested a less than 3% capacity factor for most of the turbines tested and again with some installations using more power than generated [3]. The same study also tested turbines which were mounted in rural locations and the results suggest a 19% average capacity factor, which is six to seven times more than the average of the urban based turbines [3].

Both of these studies suggest poor performance from wind turbines installed in urban locations.

1.2 INTRODUCTION

(20)

Chapter 1 Project Scope _____________________________________________________________ procedure and/or concept of a wind turbine operating in the urban environment needs to be reconsidered. To prove this idea, a 3 metre diameter HAWT (Horizontal Axis Wind Turbine) was monitored (in relatively strong and gusty winds) to review the period of which the turbine operated at its designed TSR (Tip Speed Ratio). These results can be seen in Figure 2, which clearly shows the small period when the turbine operates at the designed TSR.

Figure 2: Operating TSR of a 3m HAWT in Gusty Wind Conditions

With this in mind, it is clear that wind turbines which will operate in a gusty (urban) environment should be designed for these conditions. This dissertation looks at the theory of a SVP (Segmented Variable Pitch) wind turbine which has a possibility of working well in such conditions.

1.3 PROBLEM STATEMENT

To develop a wind turbine (based on the SVP wind turbine concept proposed by Dr Russell Phillips, and patented by NMMU) able to yield more power in gusty (urban) wind conditions than a conventional fixed blade HAWT.

1.4 SUB-PROBLEMS

The first step is to develop an aerofoil system which will allow for passive

control of the angle of attack. This aerofoil design will be the basis for each segment of the turbine blades.

(21)

The second step is to develop a scaled SVP wind turbine based on the

aerofoil system concept developed in the first step.

The third step is to develop a control system which could provide active

aerodynamic control for the pitching of each blade segment.

The fourth step is to manufacture a 3 metre diameter turbine as a

proof-of-concept.

1.5 HYPOTHESIS

A SVP wind turbine, when compared to a fixed blade wind turbine, will produce more power in gusty (urban) wind conditions due to each segment independently adjusting itself to the local wind conditions.

1.6 DELIMITATIONS

The following limitations were identified based on assumed design methods and available equipment:

 The research was focused primarily on the proof-of-concept of a segmented variable pitch wind turbine and whether the concept holds any weight with regards to increased energy yield in an urban environment.

 Calculations for the turbine blades were based on BEM theory and no tip, root, or discontinuity (between segments) losses were included in the calculations.

 The effect of rotor noise was not investigated.

 A two-blade wind turbine was the basis for the tested turbine configurations since this would simplify the manufacturing (specifically the hub and shaft) and also decrease the amount of blade sections needed to be manufactured for testing.

1.7 ASSUMPTIONS

The following assumptions were made based on design methods and available equipment:

(22)

 BEM theory was assumed to be accurate, independent of a variable pitch aerofoil (The calculated properties required from BEM theory were calculated first and then the variable pitch aerofoil properties were matched to the BEM theory results).

 The blade loading and centrifugal loads applied on the models did not affect the geometry of the turbine in any way.

1.8 SIGNIFICANCE OF THE RESEARCH

As it is with any source of power production, one of the primary considerations is cost per energy yield. A SVP wind turbine has the potential to yield more energy (in gusty conditions) for its size, therefore potentially decreasing the cost per energy yield. Since the blades are more complex this means higher cost per blade, but due to the characteristics of SVP wind turbine these costs could be recovered. These advantageous characteristics could include the following:

 No MPPT (Maximum Power Point Tracker) may be needed.

 Damaged blade sections could be fixed by replacing a single segment instead of the entire blade.

 No speed control features are required (such as pitching blades or mechanical brakes) as this is generic in the SVP design.

 The same turbine could be used for both high and low TSR requirements. Overall, there is much potential for decreasing the cost per energy yield for small wind turbines operating in an urban environment.

The concept of being able to control the angle of attack of each segment, while each segment 'adapts' to localized conditions, lends itself to various advantages over a fixed blade wind turbine.

These advantages could include the following:

 Higher power yield in gusty conditions.

 Absolute pitch control from neutral to maximum lift independently for all segments of the turbine, allowing control of power as well as the ability to

(23)

 Variable Tip Speed Ratio (TSR) which has the following benefits:

 High starting torque (low cut-in wind speed).

 Constant turbine speed for various wind speeds allowing constant voltage output therefore reducing the need for a Maximum Power Point Tracker (MPPT).

 The same turbine can be used for different purposes such as direct mechanical water pumping (high torque-low TSR) as well as effective electricity generation (low torque-high TSR).

(24)

Chapter 2 Concept Introduction _____________________________________________________________ CHAPTER 2: CONCEPT INTRODUCTION

CHAPTER 2 CONCEPT INTRODUCTION

2.1 INTRODUCTION TO PROPOSED CONCEPT

A concept has been proposed by Dr Russell Phillips and patented by NMMU for this task of harnessing wind power in the urban (gusty) environment. The concept is of a Segmented Variable Pitch (SVP) wind turbine – a turbine blade, segmented along the length, with each segment rotating freely around its pitching axis, but held at a set angle of attack due to its aerodynamic configuration as shown in Figure 3.

Figure 3: Concept Rendering of SVP Wind Turbine

Each segment will consist of two primary parts (Figure 4): the main lift producing aerofoil, and a smaller controllable aerofoil used to control the angle of attack of the main foil by controlling the pitch of the entire segment (acting as a pitch rudder). This concept allows each segment to set its own angle of attack to the varying local wind conditions (varying local air flow) at that segment of the blade.

(25)

Chapter 2 Concept Introduction _____________________________________________________________

Figure 4: Concept Diagram

Figure 5 shows a CAD model of the initial test of concept prototype built (Ø380mm). In this figure the pitch of each segment is set to an approximate angle of attack as would be while operating. Although this prototype achieved a low efficiency it did prove the concept to work.

(26)

Chapter 2 Concept Introduction _____________________________________________________________ 2.2 ALTERNATE CONCEPTS

The proposed concept of an aerodynamically stable (around a pitching axis) aerofoil system is not limited to that of having a tailed aerofoil as in Figure 4. The fundamentals for the tailed aerofoil to produce lift while aerodynamically maintaining a stable angle of attack can be reproduced in other concepts. One of these alternate concepts is based on an aerofoil profile similar to that of a flying wing (Figure 6) which uses a reflexed camber aerofoil. If designed correctly, this aerofoil will produce lift as well as provide a stable pitching moment and therefore maintain a set angle of attack relative to the air flow. This alternate concept can be implemented into the original segmented tailed concept by replacing the tailed segments with segments of this alternate concept. This new system will have similar characteristics to the original concept.

(27)

Chapter 2 Concept Introduction _____________________________________________________________ Another alternate concept to the tailed configuration is a canard design (with the small control aerofoil ahead of the main aerofoil) as shown on the right in Figure 7.

Figure 7: Tailed and Canard Test Models

A benefit of this canard design would be that the lift produced by the control aerofoil (to create a positive moment) would assist the lift of the main aerofoil. Another benefit would be that the control aerofoil would create a slot effect and accelerate the air flow over the upper surface of the main aerofoil, again producing more lift from the main aerofoil. During the initial tests of these models (Figure 7) the canard design was biased to two aerodynamically stable positions and so the tailed aerofoil concept was favoured and no further research conducted on the canard concept.

2.3 CONCLUSION

The two concepts to be tested and further researched are the tailed aerofoil and the reflexed camber aerofoil. In Chapter 3 the details of these two configurations are selected and specified. Also in Chapter 3, wind tunnel tests were conducted on both concepts and compared to the results of a common fixed aerofoil.

(28)

Chapter 3 Aerofoil Concept Theory and Results _____________________________________________________________

CHAPTER 3: AEROFOIL CONCEPT THEORY AND RESULTS

CHAPTER 3 AEROFOIL CONCEPT THEORY AND RESULTS

3.1 INTRODUCTION

In order to better understand the characteristics of the two concepts as proposed in Chapter 2, these two aerofoil concepts were tested in a wind tunnel and compared to a standard asymmetrical aerofoil. The purpose of this chapter is to obtain the properties of the aerofoil concepts, which can then be used in the design and theory of the turbine blades. Another purpose for these tests is to develop a better understanding of the different concept‟s aerodynamic stability and to realise possible manufacturing challenges.

3.2 DEVELOPMENT OF EXPERIMENTAL PLATFORM

A non-recirculating wind tunnel with a cross section of 480x480mm was used to perform the tests. Figure 8 shows a CAD model of this wind tunnel setup including the force balance and a test subject in position for testing.

Figure 8: Wind Tunnel Setup

The force balance was used to measure the lift and drag of the test subject (Figure 9). The load cells used in the force balance were two IBM

(29)

1-S2M/50N-Chapter 3 Aerofoil Concept Theory and Results _____________________________________________________________ 1 S2M tensile/compressive force transducers [7]. The data from these load cells was logged using a SoMat eDAQ (Model# eDAQ PLU8.12 S#2081) [8].

Figure 9: Force Balance Diagram

A gauge was used as shown in Figure 10 to either set (or read off) the angle of attack of the test subject.

Figure 10: Gauge for Setting/Reading the Angle of Attack

Before running the batch of tests the force balance was calibrated using a calibration mass. Also the pressure sensor used to measure the wind speed in

(30)

Chapter 3 Aerofoil Concept Theory and Results _____________________________________________________________ Figure 11 represents the test data from a sample run of an aerofoil in the wind tunnel. The test shows a consistent reading with a standard deviation of lift of 0.10N, which is less than 5% of the average reading of 2.09N.

Figure 11: Aerofoil Results (Open-Loop Wind Tunnel) 3.3 AEROFOIL CONCEPTS

Three aerofoil configurations were chosen to compare the results relatively. The first aerofoil configuration is a NACA4412 profile [9, 10] as shown in Figure 12. This profile represents a common asymmetrical aerofoil which could be used in the design of a fixed blade wind turbine. The performance results from this aerofoil will be the benchmark for the results from the other aerofoil configurations, and will also be used to validate the wind tunnel results by comparing them to the XFOIL predicted results (A reliable analysis and design software for low Reynolds number aerofoil characterization [11-14]).

(31)

Chapter 3 Aerofoil Concept Theory and Results _____________________________________________________________ The second aerofoil configuration is that of a NACA0012 profile [10, 15] with a smaller NACA0012 profile acting as a stabilizing tail as shown in Figure 13. This tail aerofoil will control the angle of attack of the main aerofoil since the entire configuration is free to pivot around a pitching axis and the tail aerofoil will act as a pitch rudder.

Figure 13: NACA 0012 Aerofoil Configuration

The third aerofoil configuration is an AOF139 profile [16] which has a reflex camber and positive pitching moment. If pivoted at the correct pitching axis location, this aerofoil will remain stable (w.r.t. pitching) and should maintain a set angle of attack.

Figure 14: AOF139 Aerofoil 3.4 AEROFOIL THEORY

For all tested concepts the main aerofoil chord length was 70mm. This chord length was the used to calculate the Reynolds numbers as in Equation 1, as well as the aerodynamic coefficients as in Equation 2 and 3 for all tested aerofoil configurations.

[1]

(32)

[3]

3.4.1. NACA4412 Theory

Since this aerofoil is well documented and could be easily simulated in XFOIL, it was chosen as the base line asymmetrical aerofoil.

3.4.2. Tailed NACA0012 Theory

For the tailed aerofoil concept as shown in Figure 13, the initial calculations assumed no aerodynamic interaction between the two aerofoils:

To simplify the design theory (as to limit the variables) the following was assumed for the chord lengths of the main and control aerofoils:

[4]

Because the segment is freely pivoted (pitching) no external pitching moment can be applied to the foil configuration and therefore for a stable system the following applies:

∑ [5]

Also since the tail aerofoil is symmetrical (NACA0012), the aerodynamic moment for this aerofoil can be assumed to be negligible, and therefore:

[6]

And based on Figure 15 and Equation 7, the stability of the system can be calculated using the sum of all the aerodynamic moments created around the pivot point: ( ) ( ) ( ) ( ) [7] where, [8] [9]

(33)

Figure 15: Diagram of NACA0012 Aerofoil Configuration

Figure 16 compares the sum of the moments around a pivot point for two systems with similarly located and angled tail aerofoils. The first system has a NACA0012 as a main aerofoil profile while the second system has a NACA4412 profile as the main aerofoil. From Figure 16 it can be observed that the x-axis intercepts represent the angle of attacks under stable conditions (where the sum of the moments are zero). It is also worth noting that since the graphs have a negative slope, the systems are stable (since a decreasing angle of attack leads to an increasing positive moment on the system). The magnitude of this slope is proportional to the stability of the system. Since the NACA4412 aerofoil has a negative coefficient of moment this decreases the stable angle of attack of the system (which can be observed in Figure 16) and therefore the control aerofoil has to work against this moment. Therefore the system based on the NACA0012 main aerofoil shows to be favourable as it is more stable and has a higher lift to drag ratio when compared to the NACA4412 main aerofoil system.

(34)

Figure 16: Stability Graph

3.4.3. AOF139 Theory

This aerofoil was originally designed for a helicopter fenestron. Its coefficient of moments (generated from XFOIL) can be seen in Figure 17 and are shown to have a generally decreasing slope with a positive coefficient of moment at a 0° angle of attack. These are the characteristics required for the theory of a SVP turbine, although not all the graphs reach a stable condition since some graphs stop short of crossing the x-axis. The graphs (coefficient of moments) however can be manipulated by shifting the “aerodynamic centre” (pivot point) and by using the sum of the moment coefficients generated from lift and drag coefficients acting around the pivot point (the new “aerodynamic centre”). Equation 10 shows this shifting of the pivot point to manipulate the moment coefficient (Figure 18).

( ) ( ) _{( )} [10]

Where „x‟ is the x-direction offset and „y‟ is the y-direction offset of the pivot

(35)

Figure 17: AOF139 Coefficient of Moments

In Figure 18 the graphs have been adjusted (relative to Figure 17) for moving the pivot point forward and show a stable system with most graphs intersecting the x-axis at approximately 9° with the exception at a Reynolds number of 50 000 where the x intercept is approximately 6°.

Figure 18: Adjusted Coefficient of Moment (Pivot Point Foreword of Aerodynamic Centre)

From the predicted stability angles in Figure 18 it can be shown that this should ensure maximum lift to drag ratios at all Reynolds numbers as shown

(36)

Figure 19: Lift to Drag Ratios at the Predicted Stable Points 3.5 WIND TUNNEL TESTING

All three concepts were tested in a wind tunnel under similar conditions.

3.5.1 NACA4412 Testing

Figure 20 shows the test configuration for the NACA4412 aerofoil. The aerofoil was fixed at a set angle of attack and the data (lift and drag) was logged for that angle. This process was repeated for all angles of attack.

(37)

3.5.2 Tailed NACA0012 Testing

The tailed NACA0012 aerofoil configuration [Figure 13] can be seen in Figure 21 as during the wind tunnel tests. During these tests the aerofoil configuration was neutrally balanced around the pivot axis and the system was allowed to pivot freely during the tests. The set angle of the tail aerofoil was adjusted for each set of data and this determined the angle of attack for the aerofoil configuration, which was observed on the angle gauge.

Figure 21: Test NACA0012 Aerofoil Configuration in Wind Tunnel

3.5.3 AOF139 Testing

The AOF139 aerofoil (Figure 14) can be seen in Figure 22 as during the wind tunnel tests at a defined angle of attack. A set of data (lift and drag) was recorded for each angle of attack.

(38)

Figure 22: Test AOF139 Aerofoil in Wind Tunnel

A second set of tests were completed on the AOF139 aerofoil in a similar procedure to the tailed NACA0012 aerofoil; the aerofoil was neutrally balanced around its pivot axis and allowed to pitch freely during the test. Figure 23 shows the shaft locating template attached (on the left) and removed (on the right). These templates were used to locate the pivot shaft and therefore shift the position of the pivot axis. The shifting of the pivot axis adjusted the “stable” angle of attack.

(39)

Chapter 3 Aerofoil Concept Theory and Results _____________________________________________________________ 3.6 AEROFOIL RESULTS

The wind tunnel results of all aerofoil concepts were recorded and evaluated.

3.6.1 NACA4412 Results

Figure 24: CAD Model of Tested NACA4412 Aerofoil

Figure 25 shows the wind tunnel results for the aerodynamic coefficients (at set angles of attack) of the NACA4412 aerofoil (Figure 24), and also include the predicted XFOIL results. The correlations between the predicted and experimental results are good and are shown to be similar.

Figure 25: Tested NACA 4412 Aerodynamic Coefficients at Re=50 000

The lift to drag ratios of the NACA4412 aerofoil are shown in Figure 26. The maximum lift to drag ratio is shown to be approximately 24.5 at an angle of

(40)

Figure 26: Tested NACA 4412 Lift to drag Ratios at Re=50 000

3.6.2 Tailed NACA0012 Results

Figure 27: CAD Model of Test NACA 0012 Aerofoil Configuration

The aerodynamic coefficients vs. angle of attack can be seen in Figure 28 for the tested tailed NACA0012 aerofoil (Figure 27). The maximum lift coefficient is shown to be approximately 0.75 which is approximately half of that of the NACA4412 aerofoil at the same Reynolds number of 50 000.

(41)

Figure 28: Tested NACA 0012 Configuration Aerodynamic Coefficients at Re=50 000

The maximum lift to drag ratio as shown in Figure 29 is approximately 12.3 which is about half of that of the NACA4412 aerofoil results.

Figure 29: Tested NACA 0012 Configuration Lift/Drag Ratios at Re=50 000 The relationship between the tail aerofoil set angle and the angle of attack of the aerofoil configuration is shown in Figure 30. The relationship is shown to be approximately linear but a change in tail angle does not result in an equal change in angle of attack. Therefore it can be assumed (since the pivot point

(42)

Chapter 3 Aerofoil Concept Theory and Results _____________________________________________________________ is through the aerodynamic centre of the main aerofoil) that the direction of the air flowing over the tail aerofoil is effected by the main aerofoil.

Figure 30: Tested NACA 0012 Tail Angle vs. Angle of Attack at Re=50 000

The aerodynamic coefficients relative to the set angle of the tail (control) aerofoil are shown in Figure 31 with a maximum coefficient of lift at -4° (which corresponds to an angle of attack of approximately 9°).

Figure 31: Tested NACA 0012 Configuration Aerodynamic Coefficients Vs. Tail Angle at Re=50 000

(43)

Figure 32: Tested NACA 0012 Configuration Lift/Drag Ratios Vs. Tail Angle at Re=50 000

3.6.3 AOF139 Results

Figure 33: CAD Model of Tested AOF139 Aerofoil

The aerodynamic coefficients vs. angle of attack for the AOF139 aerofoil (Figure 33) can be seen in Figure 34 and include the predicted XFOIL values. The correlations between the predicted and experimental results are good and are shown to be similar. The maximum lift coefficient is shown to be approximately 0.7 which is about half of that of the NACA4412 aerofoil and similar to the tailed NACA0012 aerofoil at the same Reynolds number of 50 000. However, using XFOIL to predict the maximum lift coefficients at a higher Reynolds number of 500 000 shows 1.32 for an AOF139 aerofoil which

(44)

Chapter 3 Aerofoil Concept Theory and Results _____________________________________________________________ 500 000), and so it can be predicted that the performance of this AOF139 aerofoil will improve at higher Reynolds numbers. Figure 34 also shows four sets of data (black dots) which represent a “free to pivot” system. During these “free to pivot” system tests the aerofoil was neutrally balanced around the pivot axis and the system was allowed to pivot freely. The position of the pivot was adjusted relative to the aerofoil in the x-direction for each of the four sets of data and this determined the angle of attack for the aerofoil configuration (See Figure 37). It can be seen that three of these points (from the left) coincide with the measured results of the set angle of attack tests. This shows that the “free to pivot” aerofoil is stable (about the pitching axis) and produces similar results to the fixed angle tests as can be expected.

Figure 34: Tested AOF139 Aerodynamic Coefficients (Including 4 Unfixed Test Results) at Re=50 000

Note that the black point at 9.5˚ (on the right in Figure 34) does not correlate with the fixed angle results since the “free to pivot” aerofoil was not stable at this pivot position and therefore did not produce favourable results. These unstable results could be observed during testing with obvious pitching oscillations, and can also be observed from the results shown in Figure 35 which relate the aerodynamic coefficients of the unstable system to a similar stable system. This comparative stable system is the same aerofoil but is fixed at a similar average angle of attack.

(45)

Figure 35: Tested AOF139 Stable and Unstable Aerodynamic Coefficients Vs. Lapsed Time at Re=50 000

In Figure 36 it can be seen that the lift to drag ratio is approximately 12.6 which is similar to that of the tailed NACA0012 configuration.

Figure 36: Tested AOF139 Lift/Drag Ratios at Re=50 000

Figure 37 shows the correlation between the angle of attack of the AOF139 aerofoil relative to the position of the pitching axis.

(46)

Figure 37: Tested AOF139 Angle of Attack vs. Pivot Point Position at Re=50 000 3.7 CONCLUSION

The wind tunnel tests on the two proposed concepts show that it is possible to produce a good lift to drag ratio while maintaining an angle of attack through an aerodynamic configuration. Both the tailed NACA0012 aerofoil and the AOF139 aerofoil could be possible aerofoil configurations for the segmented turbine concept, however by using XFOIL to predict performances at higher Reynolds numbers gives the advantage to the AOF139 aerofoil. Also the construction of the AOF139 aerofoil is simpler than the tailed NACA0012 aerofoil. Therefore, since the AOF139 aerofoil seems to be the better of the two concepts, Chapter 4 continues with research on a scaled SVP wind turbine with blade profiles based on the AOF139 aerofoil.

The stability of each concept within this chapter was not quantified during the wind tunnel tests and could be grounds for further research. Also the tests were limited to a Reynolds number of 50 000 and therefore tests at higher Reynolds numbers (up to 500 000) are also recommended for further research.

(47)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________

CHAPTER 4: SCALED MODEL THEORY AND RESULTS

CHAPTER 4 SCALED MODEL THEORY AND RESULTS

4.1 INTRODUCTION

The purpose of this chapter is to experimentally evaluate a scaled down Segmented Variable Pitch (SVP) wind turbine. This design theory for the scaled models is based on the outcomes from the concept tests in the wind tunnel (Chapter 3). A SVP turbine will be tested as well as a fixed blade turbine in order to directly compare the results. For all tested turbines the blade diameter will be 550mm and the chord lengths limited to a maximum of 70mm.

4.2 DEVELOPMENT OF EXPERIMENTAL PLATFORM

To test the scaled turbine blades a test platform was developed. This platform needed to log power output and turbine speed for a range of TSRs at a given wind speed. This data was then used to generate a power curve for each turbine.

4.2.1 Scaled Model Test Setup

A radio controlled aeroplane three phase permanent magnet motor [17] (referred to for the rest of this section as the „generator‟) was used to convert the directly applied mechanical energy produced by the wind turbine blades into electrical energy. In order to control the generator speed right down to almost a stop, and therefore only having a very small voltage to apply an electrical load, the internal resistance of the generator was used to apply the electrical load (i.e. the motor was shorted). This maximised the available load which could be applied. This setup can be seen in Figure 38.

(48)

Figure 38: Scaled Model Test Setup

The „shorting‟ of the generator was applied using MOSFETs which were controlled using a PWM signal (1kHz) generated from an Arduino Uno [18] microprocessor (Figure 39). The duty cycle of the PWM signal could be varied from 0 to 100% and this was used to apply a 0 to 100% of the „shorting‟ power.

(49)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________ The MOSFET bank to „short‟ the generator was designed for a 3-phase AC system, and the schematics can be been in Figure 40.

Figure 40: MOSFET Bank [19]

The overall setup for the system can be seen in Figure 41 which shows the Arduino Uno controlling the MOSFET bank with a PWM signal. The Arduino Uno also receives a signal from an inductive proximity sensor (see Figure 42), which measures the speed of the generator. The Arduino Uno then sends both the duty cycle and the measured generator speed data to the computer to be logged.

(50)

Figure 42: Inductive Proximity Sensor Setup [20]

4.2.2 Theory and mathematical model

Below are the assumptions made to calculate the power generated from the turbine [21] using the logged data (duty cycle, and speed) and are based on Figure 43.

The friction of the generator bearings can be assumed to be near zero, therefore:

[11]

The electrical internal resistance of the generator is constant:

(51)

Figure 43: Generator Equivalent Circuit

The voltage generated by the generator is directly proportional to the speed of the generator:

[13]

The current produced by the generator is directly proportional to the voltage and therefore directly proportional to the speed of the generator:

[14]

[15]

Therefore power is directly proportional to the square of the generator speed:

[16]

since

[17]

Therefore the electrical power generated is assumed to be a function of the square of the generator speed, and a constant (K) which is a function of the PWM duty cycle (DC):

(52)

4.2.3 Formulating K(DC)

To formulate the equation for the constant K, the electrical power is related to the mechanical power at various generator speeds and duty cycles (DC):

( ) ( ) [19]

where

( ) [20] To calculate the mechanical power absorbed by the generator, the torque and generator speed were logged for various generator speeds (powered by an electric drill) and various MOSFET bank duty cycles. The torque was calculated using an experimental setup were the generator mount was free to rotate. The torque transferred to the generator mount can be assumed to be the same as the torque transferred to the generator shaft as the friction of the bearings on the generator mount can be assumed to be zero. The generator is therefore used as an „electrical clutch‟ to control the loading between generator shaft and generator mounting. The measurement of this torque was quantified using a lever arm mounted directly to the generator mounting, and the force at the end of this lever measured using an electronic scale (Figure 44) and the torque calculated:

( ) [21]

(53)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________ The results of these tests are represented in Figure 45 where the measured power and the mathematical model (Equation 18 & 22) can be seen to be matched in shape and scale for the generator running at that various speeds and duty cycles of the MOSFET bank.

Figure 45: Measured Data and Model

The constant K as a function of the DC is represented in Figure 46 and Equation 22 below. The points on this graph (Figure 46) are the values of

K(DC) used to fit the mathematical model to the measured data (Figure 45)

and therefore used to generate the curve fitting function (Equation 22) using Microsoft Excel formula fitting function.

( )

(54)

Figure 46: Constant K as a function of Duty Cycle

4.2.4 Equipment Calibration

The equipment used to record data was validated to ensure accurate results as follows (See Appendix B for more details):

 The motor speed was verified using a hand tachometer (Figure 47) which measured an acceptable 1.3% error at 1296rpm and therefore validated the accuracy of the logged data.

(55)

 The electronic scale (as in Figure 44) was validated using a 500g mass and showed an acceptable 0.3% error.

 The PWM signal from the Arduino was verified using an oscilloscope (Figure 48) and the accuracy of the system was validated with a 2.33% error in the frequency and a non-observable error in the duty cycle.

Figure 48: Arduino PWM Signal

4.2.5 Testing procedure for scaled turbine

For the tests the Arduino Uno was programmed (See Appendix C for code) to run through the entire set of duty cycles starting from 0% and slowly moving up to 100%. During these tests the duty cycle and generator speed were measured and logged and these results were used to make a performance curve for the given generator. The test procedure is documented in Appendix D.

4.2.6 Conclusion

The test apparatus proves to work well as can be seen by the results of the tested wind turbine blades (in section 4.2 and 4.3) which show to be consistent and compare favourably to expected results from literature [2, 22-24]. This method of turbine blade measurement therefore proves to be simple

(56)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________ although the system is sufficient for comparative turbine tests, future improvements could include lower PWM frequencies (10 to 30Hz), as well as using a MOSFET driver controlling the MOSFETs at higher voltage, to overcome MOSFET turn on times (which can be assumed the reason for the non-linear K(DC) function).

4.3 FIXED BLADE

4.3.1 Introduction

The purpose of this section of the testing of a scaled fixed blade turbine was to form a baseline to compare the measured results of the scaled SVP turbine. Two concepts of a fixed blade turbine were designed and tested, both of which were based on a two blade configuration.

The first design was based on using the BEM theory to optimise the thrust of the blades with the priority on the angle of attack to produce the maximum lift to drag ratio. This caused a tapered blade as the aerofoil chord length was adjusted along the length of the blade in order to maintain the optimum thrust. The second design used BEM theory and maintained a fixed aerofoil chord length along the entire length of the blade. This caused the blade pitch to be adjusted along the length of the blade in order to maintain the optimum thrust as calculated by the BEM theory.

4.3.2 Theory

The aerofoil shape for both turbine blades was a NACA4412 with the blade loading calculated using the following BEM theory [5, 22-24]:

Using the following constants:

(57)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________ and

[23]

The blade was then sectioned into 10 parts and the BEM theory applied to each section. The following was then calculated for each section:

[24] [25] Assuming and

for an ideal turbine then

_and _[26]

√ ( ) [27]

Then the ideal turbine Cl and Re values for each segment was calculated

( ) [28]

[29]

The calculated Cl and Re values were used to look up the corresponding α angles [25]. The pitch of the blade at that section was then calculated using the following:

[30]

These results were validated with the following BEM theory calculations to check that the initial assumptions for a and a‟ were acceptable:

( ⁄ ) [31]

(58)

( )

( ) [33]

√ ( ) ( ) [34]

And then the lift and drag was calculated for each section using the corresponding CD value to the CL values previously determined and now validating [25].

Figure 49: Aerofoil Diagram

[35]

(59)

( ) [37]

( ) [38]

These values were related to the thrust and torque applied to the air:

( ) [39]

( ) [40]

The relative error was calculated using the following formulas: ∑ ∑ ∑ [41] ∑ ∑ ∑ [42]

The thrust error was about -5% which suggests a slightly conservative approach and an acceptable error. The torque error was about -25% which suggests again a conservative approach (i.e. the assumed forces on the momentum of the air was more than what was applied by the blades). This torque error of -25% was acceptable since the value of a‟ was 0.0096 at r = 0.2625 which was a low value, and a 25% difference in this value will result in minor differences to βr and Vr. Therefore the initial assumption of a‟=0 and a=0.33 was justified and the results were validated.

Table 1 shows the pitch and angles of attack as calculated for a constant chord length of 70mm (See Appendix E for more details).

Table 1: Properties of Constant Chord Fixed Blade Turbine

r [m[ 0.2625 0.2375 0.2125 0.1875 0.1625 0.1375 0.1125 0.0875 0.0625 0.0375 Cl [1] 0.4460 0.4897 0.5425 0.6071 0.6877 0.7899 0.9209 1.0865 1.2736 1.3656

α [degrees] 0.0 0.5 1.0 1.5 2.0 3.0 5.0 6.5 6.5 6.5

β [degrees] 82.1 81.3 80.3 79.1 77.5 75.5 72.6 68.6 62.4 52.9

γ [degrees] 82.1 81.8 81.3 80.6 79.5 78.5 77.6 75.1 68.9 59.4

Figure 50 shows the CAD model of the turbine blades based on the calculated values as per Table 1.

(60)

Figure 50: CAD Model of Fixed Two Blade Constant Chord Turbine

Once the components were 3D printed the surfaces were filled and/or sanded to obtain a smooth surface finish on the blade profiles. Figure 51 shows the realised components ready for testing.

Figure 51: Printed Turbine Blade Components

(61)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________ Table 2: Properties of Variable Chord Fixed Blade Turbine

r [m[ 0.2625 0.2375 0.2125 0.1875 0.1625 0.1375 0.1125 0.0875 0.0625 0.0375 Cl [1] 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.97 0.97 0.97 α [degrees] 6.5 6.5 6.5 6.5 6.5 6.5 6.5 6.5 6.5 6.5 β [degrees] 82.1 81.3 80.3 79.1 77.5 75.5 72.6 68.6 62.4 52.9 γ [degrees] 88.6 87.8 86.8 85.6 84.0 82.0 79.1 75.1 68.9 59.3 Chord 0.03 0.03 0.04 0.04 0.05 0.06 0.07 0.07 0.07 0.07

Figure 52 show the CAD model of the variable chord length turbine based on the calculated values shown in Table 2.

Figure 52: CAD Model of Fixed Two Blade Variable Chord Turbine

4.3.3 Results

The results for the fixed blade turbines are represented in Figure 53 and are based on Equation 43.

[43]

It can be seen that the more optimized variable chord turbine operates at a higher TSR when compared to the constant chord turbine, however does not perform well at lower TSRs whereas the constant chord turbine performs well.

(62)

Figure 53: Fixed Blade Turbine Power Curve at a Wind Speed of 4.5m/s

During testing it was also noted that the constant chord turbine self-started at 4.5m/s whereas the variable chord turbine required a small push to start operating. The maximum power coefficients of over 0.4 for both turbines suggest that they are designed well (when compared to the ideal theoretical Betz limit of 0.59 [5, 22-24]).

4.3.4 Conclusion

Both fixed blade turbines operated with fairly high coefficients of power (more than 0.4), but each turbine operates well only in a set range of TSRs. With this said, since the constant chord turbine has a relatively flat power coefficient curve and operates at a lower TSR, with a good cut-in wind speed, this concept (of a constant chord blade) may be a good fixed blade alternative for small turbines operating in urban environments. The flat power curve should assist in better power yield in an urban (gusty) environment where the TSR would be constantly fluctuating and the lower TSR would also assist in noise reduction compared to turbines operating at high TSRs. Also since the wind in urban areas is often buffeted by obstacle the lower cut-in wind speed would also assist in more power production.

(63)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________ 4.4 VARIABLE PITCH BLADE

4.4.1 Introduction

The purpose of this section is to document the theory and experimental results of a variable pitch turbine blade. The theory is based on a variable chord length, two blade turbine. From the wind tunnel tests on the airfoil concepts it can be seen that the obvious advantage of the NACA4412 aerofoil over the other two concepts is its much higher lift to drag ratio. However, looking at Figure 54 it can be noted that halving the lift to drag ratio has a theoretical non-proportional (and minimal) effect on the power coefficient and therefore the benefit of the concept of a SVP wind turbine could outweigh the disadvantage of the lower lift to drag ratios.

Figure 54: Lift/Drag Ratio Comparisons [23].

4.4.2 Theory

For the final concept of the SVP turbine, the AOF139 aerofoil was chosen over the tailed NACA0012 configuration for the following reasons:

 Simpler construction to manufacture.

 Simpler construction to handle high centrifugal loads.

(64)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________ Figure 55 shows a CAD scaled model used for the tests. The design theory was based on Blade Element Momentum (BEM) theory and the same blade loading applied as calculated for the fixed blade model. Results from the wind tunnel tests and XFOIL where used to assist with the predicted aerofoil properties which were required to estimate the blade loading.

Figure 55: CAD Model of Two Blade Segmented Variable Pitch Turbine

Table 3 shows the pivot points and chord lengths, as well as pitch angles, which were used to manufacture the scaled model of the SVP turbine.

Table 3: Properties of SVP Blade Turbine

r [m[ 0.2625 0.2375 0.2125 0.1875 0.1625 0.1375 0.1125 0.0875 0.0625 0.0375 α [degrees] 8.0 8.0 8.0 8.0 8.0 8.0 8.0 7.0 6.0 6.0 β [degrees] 82.1 81.3 80.3 79.1 77.5 75.5 72.6 68.6 62.4 52.9

γ [degrees] 90.1 89.3 88.3 87.1 85.5 83.5 80.6 75.6 68.4 58.9

Pivot from leading edge[% chord] 16.8 16.8 16.8 16.8 16.8 16.8 17.6 18.4 19.2 20.0

Pivot from leading edge [mm] 6.7 7.4 8.2 9.1 10.3 11.7 12.3 12.9 13.4 14.0

Cl 0.78 0.78 0.78 0.78 0.78 0.78 0.78 0.78 0.78 0.55

Chord 0.040 0.044 0.049 0.054 0.062 0.070 0.070 0.070 0.070 0.070

4.4.3 Results

The first two sets of data in Figure 56 shows the variable pitch results for two scenarios for the two blade configuration. The first scenario has all the segments free to rotate individually and produces a power coefficient of around 0.14. The second scenario has all 3 blade segments fastened to each other but allowed to rotate freely as a unit. This scenario only produces

(65)

Chapter 4 Scaled Model Theory and Results _____________________________________________________________ one. Both of these scenarios do not reach the designed TSR of 5, therefore showing the need for design refinement.

Figure 56: Variable Pitch Results (Wind velocity = 4.5m/s)

From Figure 56 it would seem that the blade aerofoils were not „biting‟ enough and therefore the blade loading too low. Therefore it was decided to add another blade of the same geometry to make up for the lack of thrust produced by the blades. Figure 57 shows a CAD model of the 3 blade configuration.

The development of a segmented variable pitch small horizontal axis wind turbine with active pitch control

ABSTRACT

TABLE OF CONTENTS

LIST OF FIGURES

LIST OF TABLES

GLOSSARY OF TERMS

ACKNOWLEDGEMENTS

AUTHOR‟S DECLARATION

CHAPTER 1

PROJECT SCOPE

CHAPTER 2

CONCEPT INTRODUCTION

CHAPTER 3

AEROFOIL CONCEPT THEORY AND RESULTS

CHAPTER 4

SCALED MODEL THEORY AND RESULTS