How to Build an Educational Wind Turbine

We are second year Product Design Engineering students. Our challenge was to design and build a turbine which when placed in the wind tunnel provided,  would generate the most power. The turbine was designed with the idea to teach high school students how a variety of different numbers of blades affects the efficiency of a turbine at varying wind speeds, hence the removable blades and duct. However if your turbine is for outdoor use then a simple conical duct will be more effective. This could be produced quickly using thin sheet plastic and superglue.

*UPDATE* After some stiff competition and mild explosions, our turbine came out 3rd and won us some of Tesco's finest bubbly. I'd make sure that your outer ring is entirely crack free, as ours blew up when spinning at its highest speed!

The duct, which directs airflow from the exit of the fan to the turbine blades, is important as it maximises the volume of air which passes through the turbine blades, and unifies the airflow.
This is the simplest part to produce, in that the materials required are the most basic and it is the only piece made without use of electronic equipment. 

If your turbine does not require a duct, move on to Step 6.

You will need:
A block of foam
Many, many newspapers
Clingfilm
Wallpaper paste
Masking tape
2 large sheets of MDF (roughly 300 x 400 mm)
White paint
Varnish
Hot glue gun

This foam block is to form a mould which we will later paper-mache around to create a hollow shell. The dimensions of this block are  450x280x280 mm. I produced this cuboid by gluing together 6 strips of foam  75 mm thick using a hot glue gun.
The shape we are going to form out of this is quite complex and I found it difficult to visualise. Therefore I found that sanding down a large shape was much easier than trying to build the completed shape out of measured strips, however more time consuming.

At one end of the block, mark the centre and draw a circle 140 mm radius. On the other end of the block, mark out a rectangle the same width as the block and 165mm high, again making sure that it is centred.

Now begin sanding. I used a large metal file, however low grit sandpaper would do the trick. Whilst sanding, you need to keep in mind that the middle band of your shape is to remain virtually untouched. This allows the two sides to merge smoothly together, as pictured. 

Whilst sanding the rectangular side, it will be the foam above and below the shape which you remove, whereas at the circular end it will be the width of the block that will be reduced, and  any corners rounded off.

In the final stages, use a high grit sandpaper to smooth the shape.

As our mould is made from porous material, we need to cover it with cling film to prevent the paper-mache casing from sticking to it. I used about half a roll of cling film for this.

We need to create as smooth a surface as possible to make sure that the inside of our duct produces minimal turbulence. The  easiest way of doing this is to go around the circumference with cling film once, overlap the edges, then cut the film and start again higher up until the whole shape (including the top and bottom surfaces) is covered. This technique prevents the ripples that appear in the film when you try and cover the shape in one go.

Now for the fun bit. Fill a bucket with 4 parts hot water and 1 part wallpaper granules (in that order, otherwise it goes lumpy as I discovered). Mix this until it forms a thick paste, then dip strips of newspaper into the paste and place them onto the duct mould. Cover the sides of the shape, making sure you go right up to the top and bottom edges,  but leave the top and bottom surfaces uncovered. Try and make the first layer of strips run in the same direction, and then on layer two make them perpendicular. Repeat for 8 layers.

As this shape is wider at one end and taller at the other, we are unable to simply pull the foam centre out. We need to cut the paper mache in half and then reattach the two halves once the foam is removed. A sharp craft knife or scalpel will work.

Once the foam mould is removed, the shell will distort. This makes it difficult to glue back together. Our method was quite experimental. We used a combination of PVA glue wooden supports, staples and metal weights. First, cover one side of a piece of MDF, roughly 100 x 150mm,  with PVA glue. Realign the two halves of paper mache, and then attach the MDF support across the incision. Staple along the entire length of the cut and then clamp or weight it until the PVA dries. Repeat for the opposite side.

Now you have a completed duct for your wind tunnel, but it is still quite fragile. To make the shape more rigid, hot glue wooden (or similar) supports around the two open ends. To find the dimensions of the support ring, I ran a tape measure around the circumference and calculated the diameter. Tape and/or clamp the paper mache to the wood to ensure a snug fit.

Next, coat the interior and exterior with 2 coats of varnish. This not only protects the paper mache from moisture and improves its rigidity, but will also reduces turbulence when the duct is in use. 

Finally: aesthetics. We decided to paint our duct a glossy white to in keep with our theme. 

We have access to a Rapid Prototype machine (or “3D printer”), so this gave us the opportunity to optimise our blade design to achieve as much power as possible.

Lift – based wind turbines are by far the most efficient type, so we decided to use an aerofoil (wing) shape used in wind turbines already, the imaginatively named FX-83-W-108. See http://worldofkrauss.com/foils/52

This aerofoil was chosen because it has a good Lift/Drag ratio of 68.785. This means that for every force it creates in drag, it creates 68.785 times more force in lift. The aerofoil also has a broad range of angles of attack in which it works, from -5 to +8 degrees. Basically this just gives us a little margin for error when we make the blades.

The first step in optimising the blade design is really to calculate how much power there is in the wind. Since our project involved a wind tunnel, we had a more or less constant wind speed. The formula is:

Wind Power = 0.5 * (air density) * (area) * (wind speed)^3

This gives power in Watts – make sure you use S.I units (i.e. metres, kilograms, seconds, etc.)

-The air density at sea level at 20 degrees C is about 1.204 kgm -3
-The area refers to the area that the turbine will occupy. For our design, this was the area of the end of our duct, i.e. pi * 0.14*0.14 = 0.0616 square metres.
-The wind speed is the speed of the air through the area the turbine will occupy. As you can see, a small increase in wind speed makes a large increase in power.

We had a wind speed of about 11 metres per second and an area of 0.0616 square metres, so this gave us the power in the wind as about 50 Watts.

Due to something called the “Betz Limit,” the maximum possible power that can be extracted from the wind by a turbine is 59.3% of this wind power. I won’t go in to the reasons here, but you can look it up if you’re really interested…

So now we’ve got our maximum possible power output as 59.3% of 50 Watts, which gives about 29 Watts.
This number assumes that the turbine is 100% efficient, which is impossible. The large white turbines you see all over the place these days manage about 75 – 85% efficiency, which is quite impressive. We’re not that good, so 50% efficiency sounds reasonable. This gives us the theoretical power output from our turbine as about 14 Watts.

The next bit is some more maths unfortunately – but this is the last bit!

What we need to do now is work out how big the blades need to be to achieve our calculated power output. This also depends on the speed we want the turbine to spin at.

The aerofoil we chose works best with an airspeed of about 22-30 metres per second (50-70 mph), so we need to make sure that the turbine will spin fast enough to allow this.

To work out the speed of the blade at a certain point, we use:

U = ω*r

- U is the speed of the blade
- ω is the rotational speed in radians per second
- r is the radius in metres.

We chose a rotational speed of 1500 rpm. To convert this to radians per second, multiply by 2*pi, and then divide by 60;
(1500 * 2 * pi)/60 = 157 radians per second

The blade tips will have a radius of 140mm from their centre of rotation (because of the size of the duct), so the tip speed will be:

U = ω*r = 157 * 0.14 = 22 metres per second

So this is how fast the blade is moving through the air perpendicular to the wind. To find the total airspeed experienced by the blade at the tip, we use Pythagoras:

Total speed = √((U^2 )+V^2)

U is the tip speed, measured earlier as 22 metres per second
V is the wind speed, calculated before as 11 metres per second

So we get a total airspeed of 24.6 metres per second at the blade tip, which is nicely in the middle of the range of optimum speeds for our aerofoil.

OK, next the big equation to get our blade area:

Blade area = Power/[ 0.5*ρ*√(U^2+V^2 )*(Cl UV-CdU^2)]

-Power is the wind turbine power we calculated before, 14 Watts
- ρ is the density of air, again about 1.204 kg per cubic metre
-V is the wind speed in metres per second – in this case 11m/s
-U is the tip speed of the blades in metres per second – in this case 22m/s
-Cl is the coefficient of lift for our aerofoil, found on the data sheet. Our aerofoil has a coefficient of lift of 1.138
-Cd is the coefficient of drag, which is 0.01654

So from the equation, we get the optimum blade area for our turbine’s speed and power output to be 0.003536 square metres.
We decided to have two blades (any more and they would be very small and fragile) so this gave us each blade area as 0.001768 square metres. Using a blade width of 2.5cm gives a blade length of about 7cm.

So now we have our theoretical power output, our turbine’s rotational speed, the number of blades we need, and the dimensions that the blades need to be. We’re almost ready to do a CAD model of the blades now – there’s just a tiny bit more maths first…

The final thing we need to work out is the angle of the blades at various points along the blade radius. This is for a couple of reasons – firstly, the aerofoil works best at an “angle of attack” of 5 degrees. This means that the blades will work best if they are tilted up by 5 degrees to the direction of air flow. The second reason is that the blades will experience airflow at different angles along the radius of the blade, as the blade is moving faster through the air at its tip than it is at the root.

To calculate the angle “α” that the blades need to be turned into wind from their direction of travel, we use:

α = 95 -  tan^(-1)(U/V)

-U is the speed of the blade at a specific radius (U = ω*r)
-V is the wind speed, always 11m/s in this case

Since our blades will be 7cm long, and have a maximum radius of 14cm, the root of the blade will be 7cm from the centre of rotation. So from root to tip, the angles are:

Radius(m) V(m/s)  U(m/s)  α(degrees)
0.07                 11     10.99       50.0
0.08                 11     12.56       46.2
0.09                 11     14.13       42.9
0.10                 11     15.70       40.0
0.11                 11     17.27       37.5
0.12                 11     18.84       35.3
0.13                 11     20.41       33.3
0.14                 11     21.98       31.6

OK, the maths is finally done, and now we can go on to the next step – modelling the blade in CAD software.

You can use the aerofoil coordinates from the website, save them as a .txt file, and then import them in to Solidworks to give the aerofoil shape. Once the coordinates are saved as a .txt file, go to insert > curve > curve through xyz points in Solidworks, and insert your aerofoil file on to one of the basic planes. Then select this plane, click on the sketch of the aerofoil, and select “convert entities.” This can then be scaled and rotated to a certain angle using the “move entities” toolbar.

Then, go to insert > reference geometry > insert planes, and insert 7 planes, each at a distance of 10mm from each other. Select each plane in turn, click on the aerofoil shape, and select “convert entities.” This will project the aerofoil on to each plane. As before, this can then be scaled (we used a scale of 2.5, to make the blade 2.5cm from leading to trailing edge) and you can also rotate the blade to the angles calculated before.

Then select “lofted boss/base,” and select all the angled aerofoil profiles. This will give you the main part of the blade!
All that is left to do now is make a “key” to allow the blade to slot in to the hub, and also a piece at the end to slot in to the outer ring. These can both be done by sketching on the appropriate planes, and using the “extrude” tool to make them 3D.

The blade is now ready for Rapid Prototyping!

After the blade has been rapid prototyped, it can be cast to make identical copies.

First of all though, the blade must be smoothed and polished. Most rapid prototype machines only print with an accuracy of about 0.25mm, so the blade will come out quite rough.

First, dip the blade in Methyl Ethyl Ketone (MEK). This will help to smooth out some of the imperfections. Then, apply a thin coat of U-POL, or other compatible filler, to fill in the roughness, and fix any jagged edges. After the filler has dried, sand the blade VERY CAREFULLY. Remember that the dimensions and smoothness of the aerofoil part are absolutely critical to it working correctly. Slight ripples, or changes to the shape of the aerofoil will drastically alter its aerodynamic performance.

Repeat the filling and sanding process until the blade is perfectly smooth, with no deep scratches. The blade can now be primed to show up any further imperfections, and the sanding/filling repeated until the blade is smooth and shiny.

The blade is now ready for casting.

To make the mould, you need to find (or make) a small box, about a centimetre or two larger than the blade in each direction.
Glue a small piece of plastic all along the leading edge side of the blade. The leading edge is the thicker side of the aerofoil section. Then glue this piece of plastic to the bottom of your box.

Then mix up some silicon moulding liquid as on the bottle instructions, and fill up the box.
When the silicon has dried, the box can be broken apart, and the blade can be carefully removed from the mould.
Now you can mix up resin to begin making copies of the blade. The proportions are usually about 1:1 resin to hardner. It doesn’t take long to set, so it must be poured in to the mould straight away. Make sure you roll the mould around to ensure that the resin reaches every part of the mould.

After about 15-20 minutes, your first blade should be ready. Do not be tempted to remove the blade too early – it might seem set enough, but the blade will still be soft, and will warp slightly, ruining all those angles you so enjoyed working out!
Repeat this process for as many blades as you like. We did 10, to make sure we had plenty to spare.

Then it’s the same process as before – filling and sanding. We used “green stuff” modelling filler to smooth out the little bubbles and imperfections created in the mould, and polished with fine grade sand paper. The blades can then be spray painted with any colour, as long as it’s gloss, to reduce friction with the air.

The blades are (finally!) finished.

Our hub was designed to be CNC milled from Perspex.

The first step is to sketch a circle of the correct diameter. In our case, this was 140mm. Then sketch a small circle in the middle as a centre hole.

Then sketch the same “key” shape from the bottom of the blade, and use this to create a circular sketch pattern. We only need two blades, but we created 8 identical sketches to allow for modification with different blades if desired.

Next, extrude the circle, and cut in the keys to the correct depth to match the blades. In ours this was 16mm. Make sure the centre hole goes all the way through.

Then find an appropriately sized piece of Perspex for CNC machining. It must be thick enough to allow a little more than the depth of the slots, so anything from about 20-30mm thick is ideal.

Once the hub is machined, you will need to drill out the centre hole and tap (thread) it. Our turbine will spin counter clockwise when viewed from the front, so the thread will need to be a left hand thread to make sure it tightens itself on to the shaft, rather than unscrewing itself! The size of the hole and tread depends on the size of the shaft you use, but we used an M10.

The cowl is important, as it directs the airflow smoothly to the blades.

To make our cowl, first we laminated together layers of MDF that were 160x160mm, to make a stack about 250mm in height. PVA glue works best for sticking it all together, but you will need to leave it clamped overnight to dry.

Next, lathe the MDF sandwich on a wood turning lathe to make the cowl shape. The diameter at the bottom is critical, so use callipers frequently to make sure you don’t lathe away too much.

Once you have the correct shape, use sand paper on the lathe to smooth out any roughness in the cowl.

Then add a small block of wood or MDF, about 2-4cm thick, on to the base of the cowl shape. This block must be less that the overall diameter of the base. This will raise up the cowl for the next stage – vacuum forming.

Dust over the MDF cowl with talcum powder. This will prevent the acrylic sticking in the vacuum forming. You can use any colour of 1-2mm thick acrylic for vacuum forming, but we used clear so that we could see the construction of the turbine once it is assembled.

Next, vacuum form the acrylic over the MDF shape. Once it has cooled, use a scalpel or sharp knife to carefully trim around the bottom. You should be left with a nice, neat cowl.

The next stage is to make the insert that will attach the acrylic cowl to your turbine.

First, draw a circle the same diameter as the base of your cowl (140mm). Draw another circle in the middle of this that is the same diameter as the turbine shaft, in our case 10mm. This will be the base when laser cut from 2mm clear acrylic. Glue an M10 nut on to the centre of this piece, making sure the hole in the nut is centred on the hole in the acrylic.

Then, laser cut another circle of a smaller diameter (about 40mm), again with a 10mm hole in the centre.
Thread the large circle on to the turbine shaft, followed by an M10 nut, the small circle, and another nut. You will then need to adjust the height of the small circle by winding the two nuts up and down. You need to get the two circles at the correct distance so that they both touch the inside of the cowl when it is placed over the top of the shaft. Then measure the distance between the circles, and cut a piece of clear plastic tube to that length, making sure it’s big enough to fit over the nut on the large circle.

Now drill four very small holes in the sides of the large circle, and drill holes to match in the vacuum formed cowl. The cowl can then be attached to the circles with pins and glue.

The outer ring surrounds the blades. This is another important part, as it helps to stop the blades flexing, and also reduces "tip vortices," a major source of drag. (Notice that many high performance aircraft have winglets to reduce this.)

The ring, like the hub and blades, can be modelled on a CAD program such as Solidworks. The CNC machine that we had access to  is too small to machine the ring, so it was produced using a laser cutter, from 4mm clear acrylic.

Draw the ring on your CAD software, making slots to match the end of the blades. Use a circular sketch pattern as with the hub to get all the slots identical, and in the right places. The top-down view of the ring can then be "printed" using a laser cutter.

You can also cut some rings with the same inner and outer circle diameters as before, but without the slots, to make an enclosed ring. 

The final thing to do is assemble all the parts for rapid prototyping, CNC machining and laser cutting on your CAD software, just to make sure everything fits together before you make it!

This is the frame which will hold everything together. 
We have chosen to use perspex for its rigidity, also its transparency gives the user a clear sight of how every part is connected.

To create these parts a series of CAD drawings have been generated, furthering into a CNC machine for manufacturing. 
These solidworks files are complete with dimensions.

Before the material is machined the basic shape of each component is to be cut to length,width & height, ready for the CNC machine.

Once this is done is it time to drill and thread the holes to fasten to frame. 
The best way to score accuracy is to start by clamping the whole frame together.
Once this is done you can start by drilling out the 8 holes from the pillars to the supports.

The way I achieved this is to place a 5mm drill piece(the size of the hole) in the drill. Line up the hole with the drill piece, clamp the unit to the pillar drill. Then once the drill hole is perfectly aligned, change the drill piece to 4mm(1mm smaller ready for 5mm thread) and drill 20mm into the material.

Repeat this process for the 4 holes from the base into the pillars. Where you start with a 8mm, then move down to a 7mm piece.

Once this is done you can start threading the holes. You will need a m6 & m8 tap.
Place the support in a vice, spray the holes with coolant and tap with the m6.
Repeat for the pillars with use of the m8 tap.

Now find eight 6mm bolts & four 8mm bolts to fasten then frame together.