Building a Waterwheel
Building A Waterwheel
The story of the construction of a DIY waterwheel in the UKhome > hydro | education
If you are fortunate enough to have a stream running through your land then you are blessed with a source of clean green renewable energy. In this article we will follow the construction of a complete DIY waterwheel power generation system.
Pete from Mawdesley in West Lancashire is an electronics engineer with good hands on ability with anything mechanical. With an interest in self-sufficiency, and a property with a stream running through it, the only logical thing to do was to build a waterwheel to generate electricity for his home office.
The stream crossing Pete's property has a total head of 1.5 metres. However, to protect the neighbours from the noise of a waterwheel, there is 0.5 metres of usable head. He has estimated flow rates at just 1-2 litres per second (l/s) in the summer, 40-60 l/s in the autumn/winter/spring typically, and a whopping 200-1000 l/s when the stream is in flood after heavy rain.
Damming the Stream
First of all a dam had to be constructed in order to funnel the stream toward the waterwheel. (Where damming is not possible, a Run of River set up would have been used.) The dam created a 24 inches head of water which passed through a gap equal in width to that planned for the waterwheel paddles - 2 feet.
During construction of the dam, two submersible pumps were used to divert water around a temporary 'mud' dam with a sump hole in the front. The dam was built during the summer when the flow rate was at its lowest, and so the pumps were only on 5% of the time during the 24 hour period required to get everything built.
Constructing the Waterwheel
The waterwheel itself was constructed from 12mm marine plywood which typically comes in 8ft x 4ft sheets. It was these dimensions which dictated the final diameter of the waterwheel (1.2m = 4 feet), and the width of the paddles (0.6m = 2 feet, i.e. the size obtained when a plywood sheet is cut in half). Just two sheets were required for the complete waterwheel at a cost of around £30 per sheet.
The waterwheel is made up of 2 disks (the sides), and 12 paddles. Using 12 paddles makes it easy to mark out the location of each paddle on the wheel since they are separated by a 30 degree angle (= 360 degrees / 12). The paddles were chamfered to minimise the surface area of water hit by the wheel as it rotates, with the aim of making the waterwheel as quiet as possible in operation.
Above is an image which shows how the paddles were fixed to the disk of the waterwheel. Screws were used to hold everything together, but nuts and bolts may have been better to reduce the chance of structural failure. To add strength to the whole waterwheel, 8mm metal rods were fitted underneath every paddle. With all the paddles screwed in place and everything sealed, the basic waterwheel looked like this:
Even marine ply needs to be waterproofed if it is to last any length of time without rotting in water, so two coats of Thompsons Waterseal were used to seal the plywood. This was not really enough since the ply remains porous - so much so that when the water level drops and the waterwheel stops spinning, the bottom of the wheel absorbs water where it dips in the stream. The wheel then turns unevenly for a week or so due to the uneven weight distribution which greatly increases the wear on the bearings and is less efficient. That said, after one year of operation the plywood still looks as good as new.
The FrameworkAny waterwheel needs a framework supporting an axle on which it sits rotating freely as the flowing water pushes it around. Lengths of brand new kiln dried 3" x 2" oak were used (obtained for free, otherwise a less expensive timber would have been used). The finished frame was bolted into the bed rock with 6 inch long M8 bolts as shown below during construction:
20mm bore pillow block self lube bearings (pictured below) were used for the axle (cost around £3.50 each from eBay UK - click here to search eBay for pillow block bearings).
Permanent Magnet AlternatorThe permanent magnet alternator chosen for this project was the WindBlue DC-540 - an alternator typically used in low wind areas on wind turbines, but also well suited to hydro power applications (with suitable gearing). The alternator cost just over US$200 plus shipping to the UK (which took one month).
This alternator reaches 12 Volts output at just 150 RPM, and will charge a 12 Volt battery at 15 Amps at 2000 RPM. The manufacturers claim that this unit can "handle over 10,000 RPM with ease".
Since this waterwheel rotates at just 10 RPM, a roller chain and hardened sprockets were used to gear up the waterwheel output from 10 to 275 RPM (in stages of 1:4.5 and then 1:6 to give total gearing of around 1:25).
Smaller (12mm bore) pillow block bearings (pictured below - bearings are blue) were used to hold the sprockets in place.
Waterwheel VideoClick below to watch the finished waterwheel in action after a good night's rain - generating 5 Amps at 80 Volts = 400 Watts.
Using the PowerThe WindBlue alternator comes with a built in bridge rectifier to turn the generated AC electricity into DC which can be used to charge batteries etc. Connections to the three phase AC output of the alternator are also provided so that an alternative (external) bridge rectifier can be used.
The 12 Volt battery bank consists of four 55Ah AGM deep cycle solar batteries connected in parallel for a total capacity of 220Ah at 12V. Currently 2.5mm equipment wire (multi-stranded copper) is used to connect the rectified DC output from the alternator to the battery bank 15 metres away. Since there is some voltage drop* through this relatively thin cable over the 15m run (a few percent), it will soon be replaced with thicker (10mm diameter) cable.
* Click here to try our automatic line losses calculator.
A Xantrex c35 controller (available from US$100+ new) is currently used as a charge controller to protect the batteries from overcharging. When the set battery full voltage is reached, the controller is configured to make an open circuit leaving the waterwheel (which then has no load on it) to free wheel. This could easily result in the waterwheel being damaged, and so a 12/24V resistive heater will soon be added as a dump load to be automatically turned on by the controller when the batteries are full.
The generated electricity is split between a battery in the garage and the main battery bank (pictured above) which is located in a garden office (converted shed). The charge is split through a diode and power resistor (5 x 0.1 Ohm 10 Watt rated connected in series) with 0.25-1.25A typically going into the garage battery, and 2A-6A into the office battery bank. (There is also a 65W PV solar panel fitted to the roof of the house to keep the main battery bank charged during the summer).
In the office a 600 Watt modified sine wave inverter provides 230 VAC to power a laptop and TFT display. The battery in the garage is connected to a 200W pure sine wave inverter to power lighting.
(note that a modified sine wave inverter cannot typically be used with energy efficient CFL light bulbs).
Future Improvements PlannedSince the alternator generates high voltages at low RPM it would be more efficient to reconfigure the battery bank to 24V or even 48V. This would also reduce the thickness of cable required between the alternator and battery bank, but would necessitate the purchase of (more expensive) 24V or 48V power inverter.
More Waterwheel InformationWe have more information about waterwheels available in the following articles:
Calculation of Hydro Power
Introduction to Waterwheels
Electricity from Waterwheels
Article Last Modified: 20:42, 23rd Jun 2013
Comment on this ArticleIf you have any comments on this article, please email them to firstname.lastname@example.org.
|Firstly I would like to say what an interesting article this was and glad to see UK engineering still alive and kicking.|
I agree with Simon's comments of October 15th 2013 in that the design does appear to suffer some inefficiencies due to water influx at the sides, cavitation etc.
The project could be improved as Simon says as by firstly increasing the dam from the current 0.5m to the maximum available 1.5m head available .... any noise issue can be negated as stated later.
This dam height increase will allow the wheel to be lifted to allow a gap at the bottom to allow waste water to be flued away, to prevent the wheel sitting in water when static and to remove the drag generated by the bottom of the wheel having to force its way through water. It should also allow for a slightly larger diameter wheel to be employed.
Secondly the design of the waterwheel can be improved in my honest opinion as follows:-
1) Design each of the buckets to be fully encapsulated. At present the design shows that water will run through the bottom of each 'bucket' section.
This modification allows the weight of the water in each bucket to provide additional torque to the wheel during descent. Ensuring the bottom of the wheel is above the level of the lower water level removes any back pressure as stated and allows bucket water to flow out freely.
2) Reconfigure the Bucket slats so that the water hits them at (up to) 70-80 degrees as it leaves the dam channel and fills the bucket at the top of the wheel. This means the wheel rotates in the same direction as the flow of water whereas the current wheel design results in the flowing water also potentially flowing against the back of the bucket that's rotating towards it, losing energy.
This gives your wheel the maximum energy to be garnered from the water. This combines with the torque being generated from the water in the bucket due to gravity. This implies positioning the wheel so the water flows over the top of the wheel and contacts the bucket slat on the far side past the 12 o'clock position.
It's my understanding that "Over-The-Top" or "Top-Loading" Water Wheels provide the most Energy efficient types in terms of energy extraction/ conversion.
3) Since your waterflow varies quite a lot (from 2L to 40/60L upwards) I would suggest making the wheel bucket sizes such that they hold the maximum water available for the maximum flow rate you design your dam for. This means each bucket may be quite empty in summer (1L-2L), to 50-60% in autumn/winter and full during downpours. Or maximise Bucket capacity when you have say 50/60 litre flow rate with excess flow spilling through your sleuce channels.
It means calculating the maximum flow of water versus the rotating speed of your wheel to ensure you maximise the content of each bucket without overfilling them.
Having the water hit the slat at 90 degrees and raising the wheel may be an issue regarding noise but this can be greatly reduced by applying a layer of foam such as Neoprene to the slats to lessen the impact noise. Or use cheap 2nd hand Underlay etc.
As mentioned in proposal (2) having the water directed over the 12 o'clock vertical position means creating a water channel over the top of the wheel. This can be modified to adopting a partial box configuration so that the waterwheel is partly contained. This will greatly help reduce noise as well as acting as part of the overflow channelling to prevent water impeding with the sides of the wheel etc.
20th November 2013
With regard to Pete's waterwheel construction:
(I wish I had a stream in my garden!)
1 The construction seems ok but the placing in the stream appears odd. The wheel is making use of only some of the energy provided from his dam (ok it varies during the year). Might it have been better to add a flume for a short distance - thus being a bit further downstream also improving the head - and channeling all available water through it? (Easy enough to build a controlling spillway to govern that amount)
This would lead to:
2 The possibility of a more efficient wheel.
The video shows an undershot wheel which is clearly suffering from cavitation (on exit) a) from the paddles themselves and b) from the surrounding flow of water. In other words there is no clear exit for the water being used by the wheel and this makes it inefficient.
A lower position would allow at least for a breast-shot wheel, giving greater energy transfer and possibly even for a pitchback wheel, the most efficient type. The pitchback format ensures that water leaving the paddle buckets is going in the same direction as the wheel and not curling back on itself.
Similarly it might be more efficient to build training walls either side of the wheel which would a) remove the problem from inrush of the rest of the stream and b) channel the exhaust water away cleanly. This would improve the power.
I am not really in a position to talk about the electrics other than to say the idea of a double chain drive seems rather wasteful of power. The position of the gearing and the mounting of the alternator seem very vulnerable and I would think the splash zone would mean high maintenance. Again a loss of power. Maybe an extension shaft (U/J incorporated) to a discrete and covered generating set would be a help.
October 15th 2013
Related ArticlesClick one of the links below for further articles on the following relevant topics: