Abstract

The completion of this project would not have been possible without the help and assistance of my supervisor Prof. Paul Wagstaff and Kingston University lab technicians Dean Wells and Cliff Searle. Tremendous thanks must be given for all of their time and efforts in aiding with this investigation. It served as a good insight to the nature of an engineering project and the considerations which must be taken in order to stick to given deadlines.

Project Introduction

Types of Wind Turbine

The type of turbine is defined by its method of creating rotation; lift or drag based. An example of a drag based turbine is the Savonius Wind Turbine design, Fig 1. Whereas a drag powered turbine uses the pressure difference on either side of the turbines axis to force rotation, a lift powered turbine utilises pressure differences around the surface of the blade itself. All modern turbines are lift based machines as even though drag based designs have improved significantly, the limitations involved are much greater than those of lift based turbines. According to Gipe, (1995, p171) “When examining the portion of the frontal area covered by the wind turbines blades, lift devices easily produce at least 50 times more power per unit blade area than drag devices.”

Lift devices are then further defined by the orientation of their rotating axis or prop shaft; either vertical or horizontal axis. The two solutions each have their own relative advantages and disadvantages and therefore are suited for different applications.

Horizontal axis wind turbines face the wind and use the lift produced to rotate the blades perpendicular to the wind direction. The gearbox and generator are usually located in the ‘hub’ at the top of the tower.

Horizontal axis systems operate at a lower RPM but higher torque and can be more efficient than equivalently sized vertical axis systems. As they can be manufactured on a much larger scale they can produce much higher maximum power ratings also. Modern turbines incorporate a pitching device to allow the blades to reduce the lift they provide in high winds in order to prevent damage to the system without completely shutting down.

The size of the large horizontal turbines however means they require larger average wind speeds to overcome the very large inertial forces to self start. It also then makes them very expensive to produce, transport, install and maintain. This is not helped by the fact the machinery is located at the top of the high towers.

Horizontal systems must also face the wind in order to extract its energy so requires a yaw mechanism so it can reorient itself when the wind changes direction. This occurs electronically via internal motors with the larger turbines although simpler devices such a tailfin can be incorporated into smaller designs. This extra mechanical requirement only adds to the cost of installing and maintaining a large horizontal axis turbine.

Figure 2 – Vestas V164 7.0MW HAWT c/o http://www.vestas.com

Horizontal axis wind turbines therefore, such as the Vestas V164, Fig 2, are usually deployed where the wind is relatively high, constant and predictable, usually either offshore or in sparsely populated areas to minimise visual impact for local residents.

Vertical axis wind turbines operate at a higher RPM but lower torque than equivalent horizontal turbines. They can also accept wind from any direction so require no yaw mechanisms for when the wind changes direction.

As they are generally smaller they create less visual objection and rotate with less noise.

The fact the gearbox and mechanisms do not need to be as high as horizontal axis wind turbines means they are cheaper to install and maintain.

If the wind speed does exceed the maximum, unlike the horizontal, the vertical axis turbine cannot pitch down its blades and must shut down to prevent the system being damaged. It is for these reasons that vertical turbines are suited to urban areas where wind direction can be lower and more unpredictable, as well as having less impact on the people living in the area. A common vertical axis wind turbine being installed is the Quiet Revolution QR5, Fig 3.

Wind Turbine Theory

In order to complete the project successfully a solid understanding of the wind turbine principles that will affect its performance must be grasped so that they are considered during the design process.

The aim of the turbine is to convert the energy in the wind into a useful form of energy, namely electrical energy. It does the by capturing the wind and rotating to drive a generator which converts the kinetic energy into stored electrical energy. The turbine blades of lift based devices are airfoil profile shapes which utilise the lift produced to convert it to a rotational force, i.e. Torque.

The lift acts perpendicular to mean chord on a symmetrical airfoil profile. Because of this a vertical axis turbine can only utilise a certain component of this lift force due to the nature of its rotation.

Figure 4 – VAWT Blade Vector Diagram c/o http://www.traviscarrigan.com

An optimum angle of attack must be achieved so that the profile is generating enough lift although with vertical axis wind turbines, if the angle of attack is too high the drag increases to the point where the resultant torque output is reduced.

Horizontal axis turbines can operate at much higher angles of attack so that more of the lift is converted to rotational torque.

Betz Limit

The energy transfer between kinetic energy in the wind and the turbine cannot be 100% efficient. The limit of the maximum fraction of power any turbine can extract from the wind can be shown mathematically to be 16/27 – or 59.3%. This occurs when the velocity reduction ratio (b) is 1/3, Swift-Hook (2011). The velocity reduction ratio is the ratio between the relative speeds of the incoming wind to the wind leaving the turbine.

The turbine system will also inevitably not be 100% mechanically efficient so the Betz limit of 59.3% is an ideal that can never be achieved. The most efficient modern turbines operate at 35-40% efficiency.

Cyclic Loading

As it is only when the blades are travelling into the wind they create lift, and therefore rotational force, the torque output of the turbine operates cyclically. Some turbine designs, such as the Quiet Revolution QR5 vertical axis wind turbine, Fig 3, have blades which occupy a range of angular locations and create lift along its complete length. This reduces, if not eliminates, the cyclic loading effects.

Reynolds Number

The Reynolds number is a dimensionless value which represents a ratio between the inertial and viscous forces of an object moving through a fluid. As it travels through the molecules of the gas they effectively attach to the surface of the object, creating what in known as a ‘boundary layer’. This boundary layer effects the aerodynamic properties of the object as molecules outside of it react with it as they would the object itself. This boundary layer is very complex to calculate and predict. The Reynolds number serves to provide a numerical, more objective description of the surface conditions so that the behaviour of the object can be better predicted and compared in experimental conditions.

The Reynolds number is calculated as follows:

Where: ? = density of the fluid

? = dynamic viscosity of the fluid

? = kinematic viscosity

V = relative velocity of the object to the fluid

L = the length the fluid travels

The higher the Re number, the better the attached flow over the object. Wortman (1983, p76) notes however that ‘small scale model tests are of questionable value because aerodynamic performance of the blades deteriorates at low Reynolds numbers.’

Solidity

The parameter ‘solidity’ is a ratio of the combined blade area to the total swept area of the turbine. It is a critical factor for designers of wind turbines. To determine the solidity the following equation can be used, Paraschivoiu (2002, p57):

Where: n = number of blades

C = chord length of the blades

S = total swept area of the turbine

L = the length the blades

Tip Speed ratio

A crucial factor in ensuring the turbine operates at maximum output is determining the optimum tip speed ratio, ?. This is the ratio of tip speed compared with the speed of the wind which powers it. If the turbine rotates at a ? which is too small most of the wind will pass undisturbed through the structure, meaning it is not extracting as much energy as possible. On the other hand if the blades travel at a ? which is too high, the blades will be travelling through disturbed air from the previous one, lowering the aerodynamic efficiency, or even effectively acting as a wall blocking the wind from passing altogether (Manwell et al). Solidity of the blades will also have an effect on its output. Solidity is defined as the blade area divided by the total swept area of the turbine blades and the result is effectively a ratio.

Project Definition

The objective of this project is to redesign the existing vertical axis wind turbine and construct a functional model to test in the Kingston University wind tunnel. The project utilises previous investigations, most notably for the existing model components.

Previous experiments have operated the turbine with a negative angle of attack meaning direct numerical comparisons will be limited. Recommendations on the design of the turbines construction will be taken on during the design phase in order to maximise the probability of a successful outcome.

The parameters of the turbine will be defined based on basic wind turbine principles so that their effects can be observed on testing.

Recommendations for further investigations or development of the current design will also be made.

Project Aims

The primary aim of the project is to gather experimental data to analyse in order to validate the design changes.

During the design phase specific focus will be given to reducing the start up speed of the turbine as this will give the design a distinct advantage for real world applications. The method for achieving this will be to reduce the moment of inertia of the turbine by manufacturing a lighter blade design.

The testing will also be based around investigating the effect of the augmenter on turbine output.

Another parameter under investigation is the maximum achievable power of the turbine. This will be deduced by applying various loads under various wind velocities.

Project Deliverables

The deliverables associated with achieving the aims of the investigation are as follows:

A Project Planning Report New blade concept design Detailed blade design CAD model Engineering Drawings Blade manufacture Complete turbine assembly Wind tunnel analysis Seminar presentation and Poster Final Report

Background

The turbine consists of an internal rotating structure surrounded by a metal augmenter which aims to direct the airflow onto the blades in a more controlled manner. The blades are a NACA 0018 symmetrical airfoil profile and are constructed out of fibreglass with a rapid prototype core mesh. The complete turbine can be seen in Fig 5.

Figure 5 – Previous Turbine Design

The mechanism employed to allow a range of pitch angles (+5o to -10o) in the previous model can be seen below in Fig 6. The angle of attack was altered by rotating it about a pivot using a bracket with an extended slot. This was then secured in place by a small nut. It appears this was not strong enough as a shaft extending from the central spindle to the blade was also added to give extra strength to the trailing edge.

Figure 6 – Previous Pitch Fixing Mechanism

There are 2 bearings mounted into two steel crosspieces to support the turbine, which in turn is supported by a large lattice framework. Below the turbine is a torque box hub concentric with the shaft.

The augmenter consists of eight folded steel surfaces around the circumference to direct the airflow onto the incoming blades, simultaneously increasing its lift while decreasing the drag of the system as a whole by directing it away from the blades travelling out of the wind. The turbine is omni-directional meaning it should produce the same results regardless of wind direction. The upper and lower surfaces connecting the vertical surfaces are tapered toward the turbine to direct more of the available wind onto the blades.

The previous model had been completed yet was not constructed to the optimum configuration that has been described. Also the maximum wind speed had been exceeded when it was placed in the wind tunnel which led to a catastrophic failure of the model as the RPM was too high and an imbalance in the rotor mass led to vibration, causing a collision between the turbine and augmenter. The augmenter was also distorted.

Literature Review

Ben Davis (2009)

Ben Davis investigated various blade vertical lean angles to assess the optimum configuration. The conclusion was that 0o, i.e. no lean, gave the best power output results.

The absolute data and results will not be comparable for this investigation though as the blades were pitched at a negative angle of 10o.

David Camacho – Kingston University

David Camacho identified that the ratio of the blades height (h) and the diameter of their rotation (d) plays a vital role in the overall efficiency and power output of a vertical axis wind turbine. This was determined to be at its optimum at h/d = 0.625 and has been shown that any deviation from this has a detrimental effect on the results so it is this attribute which took priority during the design process.

David also investigated various vertical lean angles of the blades by using CAD and CFD to derive as above that 0o is indeed the optimum position.

Max McDougall (2010)

Max McDougall’s investigation is the latest and much of the previous structure has been salvaged in order to save time where manufacturing new components would be unnecessary. The data however is of very limited use for comparison as Max pitched the blades at -10o (with the leading edge closer to the axis of rotation than the trailing edge) – despite him quoting it to be +10o. For this reason the real data produced is redundant for the current investigation.

The start up speed for the turbine was found to be reduced from 9m/s to 4m/s when the augmenter was included although without it required a manual start by hand. The upper limit of the wind speed in this investigation was over 20m/s – the current investigation would not expect to expose the wind turbine to such high speeds.

It was also noted that the investigation should be restricted to operating the turbine at below 200rpm as deflections in the blades and prop shaft were visible at these speeds. This also is expected to be exceeded.

Design Concept Generation

Assessment

Once an understanding of the task was gathered, work began on appraising the existing structure and assessing which remaining parts could be salvaged for the rebuild. The triangular pieces which fix the blades to the central spindle would have to be recycled regardless as they were costly and complex to manufacture. This was a driving factor with regard to optimising the height to diameter ratio.

As mentioned previously the structure had failed catastrophically so deconstruction of certain elements was not as straightforward as possible. Once the frame was removed the cruciform pieces fixing the centre spindle to the supports were easily removed. The augmenter could then be removed leaving the skeleton of the turbine itself. Members that had been added to secure the blade angles had snapped leaving the remains of its shaft in the central nut securing main shaft. This subsequently had to be sawn off as the thread of the main spindle had been ruined. After cutting the thread again the remaining nut was removed and the model reduced to its component parts.

Design Concept

The decision was taken to reduce the height of the turbine blades to meet the h/d requirements for sake of time and ease of manufacture. The height of the blades was determined by the existing diameter which currently was 0.42m, hence:

To reduce the weight as far as possible it was decided to manufacture them as a hollow structure from a lightweight composite material. This led to a concept which involves producing the complete blade as two separate halves and bonding them together.

Detailed Design Blade Design

As mentioned the blade concept is to have them as light as possible and achieving this with a hollow composite structure. The final design is shown below in Fig 7. The outermost shell is 265mm long at its outer edge. The end caps are thicker than the aerodynamic surface to aid its strength and allow for tapping threaded holes for assembly. This is the same on the inner shell so that the end caps have three times the thickness of the aerodynamic surface once assembled.

Figure 7 – Blade Concept Design

The internal profile of the end caps for the inner shell must be altered from the basic NACA 0018 to allow for mating. This profile is shown in Fig 8.

Figure 8 – Internal Blade Profile

Material Selection and Manufacture

The blades were manufactured using ‘Depression Moulding’, Gay & Hoa (2007, p19). One mould and blade half was made initially so that the thickness and finish could be assessed. Producing them in this way meant the exact length required for the opposite halve could be achieved and incorporated into the opposite mould.

The material used for the blades is Kingston University stock composite of bi-dimensional woven prepreg Carbon Fibre with E650-02 42% Resin.

As the material was hand laid and with prepreg the only consideration was the angular direction when defining the construction. Having a symmetrical layup is very important as unsymmetrical laminae cause a warped final product as a result of the curing and cooling cycle. A bending-extension coupling, Bij ? 0, occurs where a force is exerted in the x and y direction of the lamina, Barbero (2011, p182). The definition of a balanced laminate is that ‘for every +? there is another at –?…; and for each 0o lamina there is a complimentary 90o lamina with the same thickness and material’ Barbero (2011, p184) – although as mentioned the material and thickness considerations is negated by the use of prepreg.

The laminar orientation is symmetrical and balanced at [0/90/0/90].

CAD Design

A CAD model was produced for the entire assembly using the SolidWorks package, in order to assess its physical parameters at the design stage, Fig XXX.

Figure 9 – Complete Turbine Model

The final turbine design has the following physical parameters:

Blade Height (m)0.265 Turbine Diameter (m)0.424 h/D0.625 Swept Area (m^2)0.112 Blade Chord (m)0.078 Blade Area (m^2)0.02067 Number of Blades3 Solidity5.519 Manufacture Blades

To manufacture the moulds a SolidWorks CAD model was made in order to be converted to an STL file which was loaded into the CNC milling machine, Fig 10, at Kingston University. The material chosen was high density foam readily available from university stock.

Figure 10 – Moulds in CNC Milling Machine

Once the moulds were completed they had to be finished by hand. This included filing down the sharp edges around the outside of the billet in order to prevent any damage to the bag in which it was to be cured. More importantly, the inner surface of the mould needed to be smoothed to an appropriate surface finish as any roughness would be shown on the aerodynamic surface of the blade as the Carbon Fibre was cured in a vacuum. This would be detrimental to the efficiency of the blade so a rough sand paper was used, followed by a fine grade.

Also in the corners the milling machine left a small radius form the tool. These were removed by hand to ensure a clean corner for mating of the two halves.

Once the finish was satisfactory, heat proof tape was laid onto the mould surface, Fig 11, to improve its smoothness and also to allow easier removal of the piece once it was cured. Special attention was given to minimise the overlapping or gaps down the span of the blade as these also would be visible on the blade surface.

The method of layup was to place the aerodynamic surface and end caps separately but with an overlap to form one solid piece.

The Carbon Fibre was removed from the freezer where it stored and allowed to defrost in its protective bag. The profiles for the surfaces and end caps were marked out at the appropriate angles to the composite orientation and labelled for order of insertion, Fig 12. The initial surface pieces were laid at the correct 265mm length, and then increased consecutively by 1mm to allow for the required overlap at the seams. The surface and end caps were laid in alternate order, also alternating orientation on each respectively for strength.

On the aerodynamic surface, four layers of Carbon Fibre were used but on each end six were applied. This then means that when the two halves are mated, each end is twelve layers thick – three times that of thickness along the blade span. This extra thickness allows depth for holes to be tapped to secure it to the turbine structure and also gives strength to allow the loads being applied to be transferred over a greater area, reducing the risk of failure.

After laying in each piece a roller was used to squeeze out any potential air pockets and while working the area was kept clean of dust or any small pieces of debris which may be added to the piece. The use of prepreg is useful for this as the resin content is higher than desired for the final part and the removal of this resin during curing helps to eliminate these. If any of these flaws in the material occurred they would create an impurity and potential weakness in the blade as it would act as a stress concentration point. Each 1% of these trapped in the laminate leads to a 7% reduction in intralaminar shear strength and any defects of 2% and above leads to a ‘significant reduction in compressive strength’, Barbero (2011, p74).

The theory of imbalanced laminae was proven as a [0, 30, 90, -30, 0] lay up was attempted initially. The piece was removed from the mould and was indeed warped, Fig 13, which was most likely caused by inaccurate application of the 30o and -30o lamina.

The completed layup was then wrapped in a breathable fabric. This will prevent any air pockets being caught in the bag under vacuum, Fig 15.

Figure 14 – Moulds Wrapped in Breathe

Figure 15 – 3 Moulds in Oven for Curing

Once the blades were laid in the mould the bag in which they were to be cured was prepared. The bag itself was cut to size and another sheet of the breather material inserted across the base for not only the mould to rest on but also for the nozzle. If this was not added the bag itself could be pulled onto itself, disabling the ability to remove all the air from the entire bag, rendering it useles.

The bag was then sealed with a double sided rubber tape, taking extra caution to lay the bag material down with no folds or creases which may allow the entrance of air during the curing process. The functionality of the bag was tested by applying the nozzle before it entered the oven for curing, Fig 14.

The entire apparatus was then placed in the oven for 1 hour at a temperature of 100oC.

Once one section was produced the locations with four layers were 0.8mm thick once cured and cooled, and the locations with 6 layers were 1.2mm thick. This meant the second mould could be made and had a length of 262.6mm, with an altered profile of the end caps to incorporate the 0.8mmm reduction on the top half of the airfoil.

Once the two halves were removed from their moulds, excess material was removed by filing it down. The excess material refers to small flanges which were created from the sheets being laid slightly over the side of the mould and also the radii on the internal corners of the longer half which may prevent the blade mating as intended when the shorter half was inserted into place. This process had to occur above an extractor fan whilst wearing a face mask as the airborne fibres created can cause severe damage to the lungs if inhaled.

To bond the pieces to form a complete aerofoil araldite was applied to where the end surfaces meet as well as along the leading and trailing edges and the whole thing clamped together while the adhesive set, Fig 1

As the leading and trailing edges had a thickness of approximately 0.8mm and therefore a very small contact area for the adhesive to bond, a few small gaps still were visible. These would have a negative impact on the aerodynamic properties of the blades so a hardening filler was used roughly and filed down once set.

Assembly

Once all the components were ready the turbine was assembled in the machine workshop. One new auxiliary part was made from the previous assembly and that is shown in Fig 17. A piece of 5mm Aluminium was fabricated and folded, and threaded holes located to support the struts which fix the trailing edge. The previous design had inserted these struts into the nut on the central spindle. This however made disassembly difficult and the solution shown should resolve this. It also means there is more length of strut locked into not only the Aluminium but a further nut is added to increase its strength, following the recommendation of McDougall (2010, p38).

It is important that all three blades are identical in weight in order to minimise the imbalance which may occur when the turbine is rotating. As the turbine will operate at high RPMs any imbalance will have a great effect. Two of the blades weighed in at 52g and one was 53g. By re-bonding the metal end caps to the blades increased their weight slightly but meant it was possible to balance them perfectly. Once the correct configuration of blades and caps was found all three blades weighed 68g.

The central shaft also had to be turned down and threaded to allow assembly for the reduce turbine height.

Testing Theory

The goal of the test is to measure the power output of the turbine at certain wind speeds. To do this, under a known load, the RPM of the turbine was measured so a value for the power produced could be calculated using:

Where: P = Power (W)

T = Torque (Nm)

? = Angular Velocity (rad/s)

The conversion of RPM to angular velocity is:

Knowing the amount of power produced by the turbine enables us to calculate a value for the coefficient of power under those conditions using, Paraschivoiu (2002, p320):

Where: ? = Air Density (kg/m3)

V0 = Wind Speed (m/s)

S = Blade Area (m2)

The pitch angle of the blades will be set at +5o as this is the angle for best L/D, Airfoil Investigation Databse – www.worldofkrauss.com.

Procedure

Two phases of testing took place in order to compare data to that of the previous design. Initially the turbine was exposed to the wind without any load and the wind velocity increased to the point where the turbine self started. This was then repeated with the augmenter added so that the effect on start up speed could be observed.

Then by applying various loads to the turbine at constant wind speeds, the maximum power for that wind speed can be found by plotting a graph of Power vs. ?. By repeating this at increasing wind speeds a curve of Maximum Cp vs. Wind Speed should identify a trend and possible suggestion of at what wind speed the absolute maximum power may be achieved.

When taking measurements of turbine RPM care must be taken to ensure the blades are not accelerating or decelerating. To ensure this the value on the Tachometer must remain constant for 20 revolutions or 30 seconds, whichever is greater.

To apply the load the test rig was assembled with a rope wrapped around the shaft of the turbine. On one end was a mass hanger, supported by a pulley and the other end was fixed to the supporting framework. By the end with masses a torque is applied to the shaft by way of friction from the rope. Using the following equations the torque being applied, and therefore the working torque of the turbine, can be calculated.

Where: T1 = Resultant resistance (N)

T2 = Tension (N)

? = Coefficient of friction

? = Angle of contact on shaft (rad)

The angle of contact on the shaft was set at 180o, which is ensured by maintaining the two lengths of tether were exactly parallel to each other, Fig 18. The coefficient of friction, ?, was found by experiment to be 0.917.

The resistance being applied to the shaft, found by resolving T1, can then be converted to an effective applied torque using the equation:

Where: T = Torque (Nm)

F = Force applied (N)

d = Shaft radius (m)

The power of the turbine under each condition can then be calculated using the equation below:

Where: P = Power (W)

T = Torque (Nm)

? = Angular Velocity (rad/s)

The graph of Torque vs. ? should be a linear function due to this equation. From this data a graph of Power vs. ? should be a polynomial curve with a maximum value occurring between maximum torque and maximum angular velocity.

The lattice framework was stiffened to minimise vibration and flutter effects when it was in operation. To do this it was first raised off from the relatively loose metal beams running parallel to the tunnel and rested on large masses. It was also weighed down by heavy objects from the wind tunnel lab, Fig 19.

Figure 19 – Lattice Framework Stiffening

The upper framework still felt unstable so it was G-clamped on both sides to the wind tunnel opening using surplus beams from the wind tunnel lab, Fig 20. These measures ensured there was minimal vibration caused by the framework itself being deflected in the wind.

Figure 20 – Lattice Framework Supports

The procedure was then repeated with the augmenter in place. The top section of the framework was removed as a whole, Fig 21 and the augmenter lifted over the top of the turbine. This allowed the turbine to remain in position and minimise any variables that might be affected. To secure the augmenter, an M10 thread was made in the lattice members so a long bolt could be screwed in. On the other side of the bracket a nut and washer were secured, Fig 22, to ensure no slippage of the augmenter which may lead to collision with the blades. This was identical at 4 locations on both top and bottom of the augmenter.

Results

The start up speed of the turbine without the augmenter was 4.8m/s and with the addition of the augmenter it was reduced to 1.1m/s.

Initially the turbine was run with no loading.

When the turbine was rotating at low wind speeds without the augmenter it did not acquire true lift and remained at a low RPM. It was not until the wind speed was raised to around 5m/s that the turbine would accelerate to a significant RPM. When the augmenter was added the turbine RPM at lower speeds was proportional to those at higher wind speeds. The results of increasing the wind velocity on RPM of the turbine are shown in Fig 23.

Figure 23 – Graph of Wind Speed vs. RPM

The wind speed at which the turbine would accelerate into lift appears to be correlated to the start up speed. When no load was applied and the turbine was in lift the relative RPMs with and without the augmenter are almost identical. It was also observed that the turbine acceleration would sharply increase once it had reached 150-165RPM.

The addition of the augmenter therefore also had a significant effect on maintaining a constant tip speed ratio at lower wind speeds because of this – displayed in Fig 24

Figure 24 – Graph of Wind Speed vs. Tip Speed Ratio

Loads were then applied at a wind speed where the turbine was in lift. The torque applied was proportional to the reduction in RPM for both cases, which was expected

Figure 25 – Graph of Torque vs. RPM without Augmenter

Figure 26 – Graph of Torque vs. RPM with Augmenter

The gradient of the trend lines serve as a good indication of the reliability of the results. All are relatively parallel, in both cases, except for at 8m/s with no augmenter. The trend line gradient exceeds that of the data at 9m/s which suggests the values found are too high. This would have been caused by anomalies or errors in the execution of the experiment. Looking at the trend line gradients for all speeds without the augmenter suggests the behaviour of the turbine is less predictable than when the augmenter is added.

Higher loads were applied than the data points shown although the turbine decelerated to zero in these cases and so has been omitted for the purpose of analysis. The complete data set can be found in Appendix 1.

Analysis

From the raw data displayed in the previous graphs, the power of the turbine could be calculated. The graph of turbine power is shown below for each case with (Fig 28) and without (Fig 27) the augmenter.

Figure 27 – Graph of Power vs. ? without Augmenter

Figure 28 – Graph of Power vs. ? with Augmenter

Polynomial trend lines have been added to predict the entire curve based on the data found. The peak of these curves indicates the optimum power and corresponding angular velocity for the investigated wind speeds. These figures were read off the graphs and then plotted as a series to compare the augmenter effect on maximum power at different wind speeds, see Fig 29.

Figure 29 – Graph of Max Power vs. Wind Speed

The augmenter appears to increase the power of the turbine at all wind speeds. The data from testing the turbine without the augmenter at 8m/s has been shown to be higher than is realistic. If this were to be repeated the maximum power at 8m/s could be expected to be around 0.8W in this configuration.

Evaluation and Discussion

There are several elements of the turbine assembly which can be improved to increase the efficiency and therefore the power output of the design.

Blades

The blades were a novel design and as a result were subject to imperfections because they are essentially the first article.

The surface of the blades after they had been cured had a very slightly rough texture. This will increase the skin friction drag when the turbine is in motion. Also there was remains of the adhesive from the tape which acted as a barrier between the piece and the mould. It was not possible to remove this although several methods were used. Thinners, ethanol and a heat gun were all used to no effect.

The process of filling the gaps post-bonding of the two halves also could have been improved as the trailing edge was not as thin as possible. This will induce extra drag as the air struggles to reform in the stream as the airfoil travels.

There was also audible rattling of auxiliary structure in the region of 550-600rpm. After examination this was found to be due to the screws inserted into the threaded holes on the blade ends. It was found that on one of the blades, the screw attaching the bracket closest to the leading edge was not locking when screwed up fully. This meant it was slightly loose and the bracket itself was vibrating. The turbine was run up to 800rpm in this state so there was no fear of failure yet it is not a desirable situation. The cause is down to the weak nature of the thread inside the composite. For future designs, a rawl plug or socket to support the screw.

Another source of inefficiency was the drag caused by strut members securing the blade pitch angle. If possible a design should be developed to secure the blades at the correct angle without these members.

Yet more inefficiency would have been caused by the spindle arm pieces which connect the blades to the shaft. From a previous experiment where the turbine failed and general wear and tear from others have left these pieces scuffed. At the extreme ends of these there are extension pieces to increase the turbine diameter. The mating of the two has left gaps which will trap air when it is rotating. Both of these factors will also increase the drag unnecessarily.

The augmenter was not altered in order to complete the project on time. The preferred option would be to shorten it down to leave approximately 10% of the blade height distance between the edge of the blade and the upper and lower cowling. Also the upper and lower rim itself would be at best effect at 30o to the incoming wind; it presently sits at around 10o. This would be a useful update for future investigations.

The testing procedure is open to scrutiny. The absolute value for power output is dependent on the coefficient of friction used between the tether and the shaft. The method for deducing this was to fix the turbine blades and place a mass hanger on two pulleys at either end of the tether. Imbalanced masses were added to the hangers and the rate of acceleration that the heavier one fell was calculated. Knowing the acceleration, a difference between that figure and 9.81m/s 2 of gravity was assumed to be caused by the friction applied at the shaft. This resistive force was converted to a friction coefficient for the materials. The timing of the falling masses was repeated 10 times and all values were in a range of 0.2 seconds. The value of 0.917 however seems high and errors may have occurred in either the timing or the height at which the masses were dropped.

This doesn’t mean the results are worthless though. As long as the same materials and hence coefficient of friction is used, which it was, then the varying results for different conditions are still valid if only as a comparison and therefore the aim of finding optimum conditions is still achieved.

A further alteration which would of much improvement to the investigation is the material used for this tether. At high applied loads then the tether ‘burned out’ as the turbine was at high RPMs also. This meant applying a new tether very often and possible discrepancies in the results. A better material or even better method, with greater accuracy would enable the absolute results to be much more reliable.

Conclusion

The addition of the augmenter not only greatly reduces the start up speed of the turbine but improves performance, especially at low wind speeds. When there is no loading and the turbine is in lift the augmenter does not affect the RPM however with no load the turbine is not useful. It is only when a load is applied that power can be extracted. The augmenter does have a significant effect on this and allows the turbine to not only operate with a higher power output but also appears to give greater predictability as a result of directing the airflow.

Recommendations

The following recommendations are made to improve the investigation in the future:

References

Gipe, Paul (1995) Wind Energy Comes of Age, John Wiley & Sons Inc, Canada

Davis, Ben (2009) Vertical Axis Wind Turbine Blades for Urban Environments, Unpublished MEng Dissertation, Kingston University, London

Perera, Kapuruge (2009) Analysis and Development of a Vertical Axis Wind Turbine, Unpublished Final Year Individual Project Report, Kingston University, London

Bhatt, Hardik (2009) Vertical Axis Wind Turbine, Unpublished Final Year Individual Project Report, Kingston University, London

McDougall, Max (2010) Vertical Axis Wind Turbine Project, Unpublished Final Year Individual Project Report, Kingston University, London

Pathmanathan, A.V (2008) Analysis on a Vertical Axis Wind Turbine, Unpublished Final Year Individual Project Report, Kingston University, London

Gay, D & Hoa, Suong. V (2007) Composite Materials Design and Applications – 2nd Edition, Taylor & Francis Group, USA

Barbero, E.J (2011) Introduction to Composite Materials Design – 2nd Edition, Taylor & Francis Group, USA

Wortman, W. J (1983) Introduction to Wind Turbine Engineering, Butterworth Publishers, USA

Paraschivoiu, I (2002) Wind Turbine Design, Ecole Polytechnique de Montreal, Canada

Airfoil Investigation Database, Available at http://www.worldofkrauss.com, accessed on 8/3/2011.