Hydroelectric Microcontroller

[JJS1999]

Hi, I'm a computer engineering student in South Africa and started my final year now. Obviously this means I have my final year project. I'm still playing around with some ideas and one of them was a hydroelectric powered microcontroller. Essentially South Africa suffers from what we call Loadshedding, where pretty much every day our power is turned off for 2 hours ON A GOOD DAY. Sometimes it can get as bad as 10 hours in a day where we're without power.

So for that purpose I was looking at renewable energy. Solar power is pretty pricey still over here and unless you're device is outside you can forget about wind power. So I thought, what about water? If you can power a motor that generates electricity with running water through tubes, that in turn powers another motor that can ensure the constant flow of water, wouldn't that ensure you're device can run for at least 4 hours (which is the longest single period of Loadshedding we currently can get)?

Thing is, whenever I google most hydroelectric generators are huge, so now I'm wondering would a small scale one actually work to power a microcontroller which can be used to do other things (what that other thing is I'm still thinking of). But this is basically where I'm starting, but I'd like to hear from other engineering students if I'm going towards a dead end or if the idea has some merit.

 

[3DDave]

There is a system called "pumped storage" where, when there is excess power producing capacity, water is pumped to a higher altitude and then, when there is excess demand, the water is allowed to flow back through a generator. Usually that altitude is several hundred feet. You can figure out the flow rate and pressure to get the power required : Power = (P x Q) ÷ 600 - where power is in kilowatts [kW], P is the pressure in bars, and Q is the flow in litres per minute.

However, for small flows the friction losses in the flow path will dominate as will losses due to turbulence in small turbines. It can work, it's just not efficient. If there is enough water, that may not be a problem. Solar power is as low as 20% conversion efficiency, but the sun makes enough it's still useful.

What won't work is to have a water supply drive a generator to drive the pump that drives the water supply. That is in the class of "perpetual motion machines."

The first step is to determine how much power you need to run the device and then how long you need to run the device. From that you calculate the amount of energy you need to store or have access to, not including losses to inefficiencies.

Rechargeable batteries are the usual approach - the efficiencies of recharging are high, the efficiencies of extracting energy is also high.

The other part of this is to examine the use of the device to see if it can use less power. Many microcontrollers can be set to rest for the majority of the time using microwatts awaiting external inputs to wake them; often the largest draws are other chips on the same board. In many devices power to peripheral chips is controlled by fet transistors, allowing the peripheral chips to be shut down completely.

 

[JJS1999]

Thank you for this explanation. I always wondered why it's never used and now I understand. I'll see if it's possible to use the pumped storage method and compare that to a rechargeable battery and see which I can to use to optimize my system most efficiently and use the best method. Thank you very much

 

[waross]

It may have been better to have done this first.

One Watt Second = 0.7376 Foot Pounds per Second.
Stored energy needed for one Watt minute = 44,26 Ft pounds
Stored energy needed for one Watt hour = 2655 Ft pounds

For 20 Watts for one hour, drop 2655 pounds through a distance of 20 feet.
Add a few pounds for losses.
If you can do it, you can do it.
(Your losses in gearing may be significant)

--------------------
Ohm's law
Not just a good idea;
It's the LAW!