Hydrogen fuel cells (in particular PEMFCs) are an exciting technology which has been around for a while now. Unfortunately, they haven't really take off (excuse the pun) due to their high cost, poor lifetime and slow transient performance.


1. So why are people still interested?


Well, firstly you have the good old "clean & green" argument. In operation, the take in oxygen (from the air) and hydrogen (the fuel), and give out heat, water and electricity. No carbon dioxide, monoxide, sulfides and other nasty things.

Secondly, they're not dependent on fossil fuels to run, and are currently the most promising energy generation technology that is independent of the elements (wind, solar, etc).

Hydrogen has a superb energy density, meaning not much weight gives you a huge endurance.

Fuel cells tend to have no moving parts, unless the have a cooling fan or air supply fan, so they are mechanically very simple.

2. So why aren't we using them?


They require platinum as a catalyst so make the chemical reactions work. Platinum is incredibly expensive, and more annoyingly it tends to be washed away over time which reduces the fuel cells lifespan.

They are slow to respond, so cruising along at 10% power, then suddenly demanding a full 100% may take anything from seconds, to minutes, to hours, depending on the type of fuel cell. This proves a big challenge in mechanical transport, such as planes, and cars.

Hydrogen is a pain in the rear, generally. To produce the high quantities we need, we can only do it (cost-effectively) by burning fossil fuels. Of course it can be generated using renewables (wind, solar, etc), but this is far too expensive at the moment for the industry to consider.

Hydrogen also is a pain to store. Because there are so few molecules in such a large amount of space (low density), you have to store it at tremendous pressures to get a usable quantity, which requires huge storage tanks. Not particularly practical in transport applications. There are other methods, but cost and practicality is still a huge issue.


3. Where do I fit in?


To start I am not a chemical engineer. However, I am a PhD researcher for the Hydrogen Fuel Cells & their Applications Doctoral Training Centre (DTC) in the United Kingdom, so I know a fair amount about it. By trade I am a autonomous control engineer, specialising in unmanned aerial vehicles (UAVs).

Given all the problems outlined above, even the suggestion of putting a fuel cell on a small UAV, which is essentially a modified off the shelf remote controlled hobby plane, seems crazy. But it is do-able, and has been done. The advantages for flight aren't the clean & green approach currently dominating the selling point of fuel cells, but that key word "endurance". To fly for 10 hours, a UAV would have too many batteries to float, let alone fly, but with hydrogen's energy density this may well be possible. That's what I'm going to find out.


4. So that's nearly 500 words with no raspberryPi?


Haha, hopefully I have set the scene now. Let's talk about the electronics!

The fuel cell is a dumb system. It's just a contained and continuous chemical reaction. But it can be controlled through the hydrogen supply, the load and the temperature. I can overcome the poor transient performance if I know that "soon" I will need full power; I can simply crank the fuel cell up, dump the excess power, then when the time comes, 100% is there to give. Sounds wasteful, but it turns this fuel cell into something useful. To control the logic we need a computer. On a small UAV the raspberryPi is ideal, in terms of size, power (in & out), flexibility and functionality.

Running a Real Time Operating System (RTOS) is a seriously useful tool, which I could not do on my work last year (using Arduino). I can programme in any language, use other peoples software (eg the camera) and share my code with the world (and myself for backup!) on GitHub.

So far I am using several add-ons including the MATLAB/SIMULINK support (to help the academics!), rPi Camera, delta-sigma-pi I2C ADC, piFace SPI Digital IO, TMP102 I2C temperature sensors and have the whole lot connected to an onboard UAV autopilot.


At risk of making this essay any longer, I'll wrap up now. I hope this is something that tickles your interest, and keep tuned for some videos, pictures tutorials and demos in the coming weeks. In the meantime, if you are interested in the plane the check out my page on diydrones.





Academic Researcher

Fuel Cells in Unmanned Vehicles

Autonomous Systems Specialist

Aeronautical & Automotive Engineering

Loughborough University