Powering the Internet of Stuff – by sucking electricity from TREES
Where are my generating wellies?
Feature Despite regular headlines about self-powered gadgets and a deluge of stories claiming that any day now we should expect our smart phones to start gathering power from the environment around us, the promise of harvested energy always seems just out of reach. Or is it?
We may not be charging our gadgets in our Wellington boots any time soon, but thousands of sensors and switches are already casting off their batteries and going it alone.
Every year, Orange, the UK mobile operator, used to announce a new technology with which to charge mobile telephones. The release always came at the start of festival season, with enough canned quotes and photo-shoots with pretty girls to appeal to the laziest of journalists.
In 2011 we brought you the news that t-shirts could collect power from sound waves, in 2010 it was wellies with a charger built into the heel, 2009 predicted tiny windmills everywhere, and in 2008 we would be charging phones through the power of interpretive dance. Yet, despite all of that, we're still lugging mains-powered chargers around the place.
The problem with energy harvesting is that there really isn't very much energy to harvest. Researchers trying to create enough power for mobile gadgets, even wearable technologies, struggle to collect energy from our movement or the air around us. Prototypes are easy, and headline-grabbing, but commercial products turn out to be ineffective, encumbering, or both.
Dance Charger: she looks so comfy and stylish too
Yet there are places where energy harvesting is much, much, easier, and more useful too. A human can usually find a plug socket, and since we all started using the same cables that's got a lot easier. Yet the Internet of Things isn't about people; it's about... er, “Things” ... and they aren't always so proactive in powering themselves.
Take Awesense, the Canadian company which makes a living scattering sensors around electrical grids to see where power is being stolen from. The sensors are hung off high-voltage supply lines, and report back over a wireless mesh, so can be fitted without interrupting the supply, while Awesense takes a cut of the value of any thefts its kit detects.
The limiting factor, on the sensors, is the life of the battery which has to constantly transmit the measured current to a receiver up to a kilometre away. Awesense is, therefore, about to announce a new sensor which can induce voltage from the monitored line, harvesting energy from the very thing being monitored.
Collecting induced current from directly alongside a high-voltage wire is obviously a lot easier than eking current out of a pair of wellies, but it sill demands careful husbanding of the available energy, as the sensor must function even if the measured current is zero. That does get easier every year as semiconductor technology improves and power consumption drops.
Awesense SenseNET Raptor: Fitted using a long pole, without de energising the wire
The ability to do more computing, with less energy, is probably as important as the developments in harvesting technology, in making the technique practical, but it's in the collection of ambient energy that the most interesting innovations have been seen.
Awesense is quite unique in having such a rich source of power so close by, though there is a surprising quantity of energy around the place once one starts looking.
The Sun has got his (old) hat on
By harvesting standards, Solar energy is old hat. Photovoltaic cells turn sunlight into electricity with limited efficiency but are perfectly adequate if the power required is minimal and the light source consistent. The best example is solar-powered calculators, which have the added advantage of being able to rely on a human to take them out of the shade.
Energy harvesting versus device power consumption. Source: Kamal Shah, Intel
In the Internet of Things solar power is less interesting, as shade can pop up from anywhere and the intermittent supply from light means that harvesting has to charge a battery (or capacitor) to provide a consistent supply.
The other traditional form of renewable energy, wind, is equally useless, as it doesn't scale well. The nature of a circle means that adding 20cm to a 2m turbine blade increases the catchable wind by almost half (4.52sq m compared to 3.14), while removing 1cm reduces the harvested wind by half a square metre*, meaning that tiny turbines are no more than novelties, despite the enormous amount of backing they keep getting on Kickstarter.
Practical energy harvesting needs an alternative supply, but fortunately there are a handful of techniques being developed.
The alternative to alternative energy
First up is piezoelectricity: a flow of electrons generated by squeezing a crystal, turning vibration or other mechanical energy into electricity. The most popular use of piezoelectricity is in sensors, such as a pickup on an acoustic guitar – the vibrations generate electricity which is amplified – but anyone who's used a barbecue lighter (or a posh cigarette lighter) will have pressed a button to generate an electrical spark which comes from a compressed crystal.
Piezoelectricity is also used by the latest generation of light switches, which use power generated from the pressure applied to send a radio signal to the computerised lighting system. Systems such as the Philips Hue lights – multicoloured bulbs controlled from a central hub – can now be wirelessly connected to an energy-harvesting switch.
The Philips switch is really four switches, allowing the user to pre-program four mood settings which can be activated without booting up a smartphone, but the power comes from the action of pressing the switch.
Philips Hue Tap: She'll have to hold it tight, as it's easier to press when on a wall
In use, the control is a little strange. We're used to switches being mechanical, or sensitive to the slightest touch, but the Hue switch clearly needs to be pressed and the firmness with which it needs to be pushed can catch out those unfamiliar with it. If the switch fails to work, the correct behaviour is to push it again – slightly slower and with more pressure – but that is a learnt behaviour and far from intuitive.
It might be a novelty in wireless light switches, where a battery would generally work equally well, but there are other areas where piezoelectricity is the perfect solution, such as monitoring the performance of heavy machinery.
Feel the vibe
Advanced Cerametrics makes wireless sensors with magnets on the back. The magnets are used to hold the sensor in place – slap them onto a big machine and they'll start transmitting performance information powered entirely by piezoelectricity, generated from the vibrations of the machine. That vibration is also the information transmitted, as changes to the frequency or amplitude can indicate developing faults, so the sensors take their power and data from the same source.
Perpetuum's Rail Solutions do much the same thing with rolling stock, and can spot worn bearings, failing gearboxes and anything else which turns. The sensors gather data while the carriage is vibrating in motion, while a capacitor stores enough charge to upload the information on arrival at the station.
Perpetuum wireless sensor node: Place in the most inaccessible of places, the sensor needs no maintenance
Southeastern Trains, the Govia-owned operator of the Kent rail franchise, has fitted Perpetuum sensors to its fleet of 618 Bombardier Electrostar carriages, convinced that early identification of faults will pay off in reduced maintenance costs.
The devices being made by Advanced Cerametrics, and its competitors, are a perfect example of how the Internet of Things is so much more than wrist watches and smart glasses, and energy harvesting is a key element in making it work. Batteries would probably be cheaper, and certainly provide more power, but having to replace them can be prohibitive when the working environment is the underside of a train, or an industrial generator.
Sensors powered by wasted energy – in this case the vibrations of the equipment – can run indefinitely without worrying about their operational life, so owners can build up a knowledge of what constitutes "normal" behaviour which makes spotting anomalies so much easier.
Yet vibrations aren't the only source of ambient energy, and piezoelectrics aren't the only way to harvest power.
Going hot and cold on the alternatives
Any temperature difference can be harnessed to create electricity, based on the Seebeck effect which utilises two metals placed side by side with conductors above and below. As the metals react differently to the heat differential between the top and the bottom, current is created. Manufacturers vie to create metal alloys which generate the greatest current in the widest range of temperatures.
Tellurex Enviro-Light: just light a candle and before you know it you'll have light enough to see!
The technique scales quite well, enabling the BioLite Stove to run a fan. That improves combustion, and charge an iPhone too, while Enviro-Light gathers the heat of a candle to power a light, which is, we have to assume, brighter than the candle whose heat it consumes. Several companies are creating much smaller thermal generators for use in less-headline-grabbing Things.
Thermo Life, for example, has a charger the size of a ha'penny which (the company claims) can put out 3 volts of electricity (at 10 microamps), while Perpetua will sell you the delightfully-named Power Puck, which uses an array of fins to create a temperature difference when magnetically affixed to something hot. The 'puck is designed to power pressure sensors from GE, Honeywell and Rosemount, but the company also sells a Power Tile which will – when slapped onto a pipe 90°C hotter than the surrounding air – supply 24 volts topping out at 20mA.
Perpertua Power: Surely the best-looking of the harvesting technologies
So attractive is the Seebeck effect that it's even available for in-home use, thanks to Micropelt: maker of the wireless radiator valve which turns boring old (manual) valves into intelligent Things powered by the heat of the water running past them. The "intelligent Thermostatic Radiator Valve" uses EnOcean's highly-efficient (but proprietary) radio protocol to integrate with other kit from EnOcean, providing truly-wireless home automation.
If a tree falls in the woods, will it still recharge my iPhone?
Such proprietary protocols permit even tiny amounts of power to be usefully harvested. The standards (ZigBee Green, Dust Networks, Weightless, etc.) are catching up, but when the power is being squeezed out of the trees then every smidgen has to be utilised to achieve something worthwhile.
Voltree: Big bit of kit, but a very small amount of power
Getting electricity out of a tree might seem a small step from getting blood from a stone, but in fact it's surprisingly easy, though using that energy is harder. A nail driven into the tree a few feet from the ground, with another in the ground nearby and a circuit between them to collect the voltage, was demonstrated by researchers at MIT in 2009. Doing anything useful with that power has been a challenge.
Voltree was born from that research, and wants to run a wireless mesh capable of detecting forest fires. The company has been running trials with the US forestry service: tree-powered sensors could, in theory, be easily deployed and forgotten about, until they get hot enough. The company also sells an educational kit for schools who want to see how much power their local trees can produce (we're talking microwatts here), with a view to understanding more about the process.
The power, apparently, comes from the pH difference between the inside of the tree and the ground. The tree has to maintain that difference to grow, so the mechanism really is tapping the tree (as opposed to drawing energy from the degradation of one of the nails, as a potato battery does) and will provide energy indefinitely - though, again, only in very small quantities.
Detecting fires is no-doubt useful, but it's not immediately obvious what other applications will be found for Voltree's technology, and the company is still struggling to turn what is has into a sustainable business as it's right at the edge of what's possible with ambient energy harvesting.
But in that respect Voltree is far from alone. There are many more companies making harvesters than companies making products which use them, for the moment at least. Numerous suppliers will deliver evaluation kits, often featuring multiple harvesting techniques, to enable the entrepreneur or inventor to find out what is possible. When Things become self-powered it opens up a huge range of possibilities, but we're only just starting to understand how to exploit them.
Wurth is one of many offering multiple techniques on a PCB: this device has solar, thermal, vibrational and magnetic energy harvesting
Energy harvesting is only one part of the IoT puzzle - printed electronics, highly efficient MOSFET chips, greater intelligence, and better radio technologies, are all playing their part in making better Things to manage our lives. Few people can reasonably justify spending ninety quid on a Nest smoke alarm just because it can talk to Google, but we can all see the value is knowing a train is going to fall apart before it happens.
Bringing it back to the Things around us
The popular press keeps banging on about smart thermostats and wearable technology, as though that were the Internet of Things, but such applications are already in the minority compared to the devices which proliferate wherever there is enough energy to run them. The Internet of Things isn't a buzz of unmitigated hype, it's just that one has to wade through a shitload of nonsense to see that the proliferation of self-powered Things really is going to change the world. ®
* Thanks to all the commentards and emailers who pointed out the original sums were wrong, stating 1cm instead of 20cm. Vulture Central's backroom gremlins have revised their very rusty knowledge of geometry and amended the story accordingly.