Self-powered Internet of Things, From Theory to Reality – Small Sensors for Large Machines

In my former posts, I discussed conditions for an Internet of Things (IoT) to become reality. Some of the aspects are still on a wish list. We will see the results only in several years. Nevertheless, there are already concrete developments that take the fast lane in the direction to a real IoT, for instance in the field of energy-harvesting wireless technology. The specific characteristics of this technology enable sensors and actuators to work with an unlimited source of energy harvested from the surrounding environment, making them suitable for many IoT applications where flexibility and maintenance-free operation is highly requested.

Strong power sources are temperature differences, for instance. A difference of just 1°C (1 K) is sufficient to power a wireless sensor. A temperature difference of >10o C allows the operation of more elaborate wireless sensors to monitor and report conditions. The secret behind is an optimized combination of a Peltier element and a DC/DC converter. Standalone Peltier elements only produce very small voltages of about 10 mV per degree Kelvin.

Electronic circuitry connected to this, a sensor module for example, needs a typical supply voltage of 3 V. Therefore, a DC/DC converter is needed to complement the system. An optimized oscillator already starts to resonate upwards of 10 mV input voltage. On 20 mV or more (i.e. about 2K), a useful output voltage of more than 3 V is generated. For a temperature difference of only 7K, approximately 100 μW of energy is generated.

This kind of energy-harvesting principle is particularly interesting for predictive maintenance applications in the industrial sector. In production plants, virtually every unit of industrial machinery generates waste heat. Here, we find perfect conditions for the use of thermal-powered sensors. Typically, the temperature difference between the machine’s heat and the cooler environment provides enough energy to realize a complete autonomous wireless sensor that measures and sends data twice a day. Some might now ask: only twice a day? Yes, this is sufficient for analyzing oscillation of bearings and for monitoring deviations from the normal status and irregularities of measured values. Such sensor systems achieve highly attractive scaling effects.

The measurement prevents unexpected failures and allows a better planning of changing the bearings. In that way, the maintenance intervals are extended, whereas malfunctions, major damage, and thus production downtimes are avoided. That is an important cost benefit, as only one hour of downtime in a paper plant would cost up to $7,000. In addition, the Total Cost of Ownership (TCO) is reduced as service staff is only required when the sensor reports a necessary servicing or repair.

Self-powered sensors positioned at machines can measure data from many different points where power cables or batteries would prove to be a drawback. Batteries last for only a limited time and must therefore be replaced regularly, which is sometimes impossible if the sensor is placed at a point where a change can only be done by stopping the machine. It is out of the question that this would be opposed to a monitoring system which is intended to avoid downtime. In addition, the initial installation time is significantly reduced. The self-powered sensors can flexibly be positioned at the machine. An IP gateway, which can be a plug-in receiver for instance, receives the encrypted wireless signal from the sensors and sends it to a monitoring PC — done. Altogether, such a functioning installation can be completed within 15 minutes or less.

Using energy-harvesting wireless technology, an intelligent predictive maintenance can be realized at affordable costs, even in retrofit projects. It enables the connection of a large number of battery-less, maintenance-free sensors into an IP network that processes data for intelligent and safe conditions monitoring and for a better understanding of technical systems. The best thing: This IoT application is already stepping into reality with field trials in real industrial environments.

18 comments on “Self-powered Internet of Things, From Theory to Reality – Small Sensors for Large Machines

  1. Davidled
    June 19, 2014

    Piezoelectric generator generates electric voltage by mechanic vibration. This is one of popular energy sources. This is currently used in the tire pressure wireless sensor. But, engineer needs to determine how much energy the wireless sensor requires averagely as well as in downtime before developing energy source.

    June 21, 2014

    Hi Mathias,

    Can such temperature difference based harvesting be used in medical device implants? For example, pacemaker application. Once installed inside the body, it can go on forever without the need for replacing the battery.


    June 21, 2014

    Hi Jayaraman, thanks for your question. While we are exploring at EnOcean the use of thermo-harvesters in conjunction with the human body, an application completely inside the body is rather difficult. Any thermo-harvester needs a temperature differential to generate energy. If you placed the device into the skin, with one side being exposed to the ambient air temperature and the other side surrounded by human body temperature the harvester would generate energy. Whether the energy generated was sufficient to power a pacemaker would need to be evaluated. Kind regards, Matthias.

  4. Netcrawl
    June 22, 2014

    @Jayaraman, currently, electronic medical implants such as pacemakers and neustimulators rely on batteries that require constant surgical replacement, progress had been made. And in the future  we could expect something new- a convenient permanent solution that could harvest the thermal energy of the human body through a simple heat engine. I think its possible. 

  5. Netcrawl
    June 22, 2014

    @Daej I agree with you about piezoelectric generator, it a has huge potential. Cardiac pacemakers, powered by piezoelectric energy harvested from the patient's heartbeat, are now viable, eliminating the battery avoiding replacement and maintenance. This also means that implants can be made smaller, more convenient, more effective and longer lasting. Saving us a huge amount of money and effort.

  6. samicksha
    June 23, 2014

    I guess one of the commonly used application is the electric cigarette lighter wherein pressing the button hits a piezoelectric crystal, producing a sufficiently voltage electric current that flows across a small spark gap, thus heating and igniting the gas.

  7. GSKrasle
    June 24, 2014

    I find this whole topic exciting because I have a specific application whose solution would serve in other places and for other purposes.

    I have a 'millivolt control' gas greenhouse heater whose pilot-light plays on a 'thermopile' that provides (500mV?) to directly power the pilot valve and (through thermostat contacts) the main burner valve. It works great, except that there are no (non-battery) 'smarts' available, no set-back or logging thermocontroller. There is also the necessity of the pilot continuously running.

    Pumped hot-water is a preferable heat-source in this application, and I do have a battery-backed-up pump and 'dumb' water-heater, but the continuously-running pilot is an energy-waster, and it would save energy if I could programmatically control the 'boiler' temperature, reducing it, or shutting it off, when heat is less needed. But I need to retain independence from the AC mains.

    While I have battery-power available that could run a smart tank thermostat and igniter (in addition to the circulation pumps), it would be more convenient to have that subsystem autonomous. Then it would be suitable to saving energy in normal residential installations.

    If the tank controller could use the tank's bulk temperature as a primary energy-source, perhaps supplemented by a thermopile on the burner for (manual) cold-starting and maybe a battery or wall-wart for back-up/automatic-start, it could have the reliability I require in this application, as well as for normal household use, so that loss of AC power would not be completely crippling.

    From my perspective, this looks like a relatively simple and high ROI project; I am surprized I can't just go out and buy it! The regulatory/safety issues around messing with the gas-supply are all that worries me against building my own.

    A similar case can be made for implementing Delta-T operated controllers and pumps for heat-accumulation applications like thermosolar (which can also use PV) and fireplace/woodstove systems: convection-based systems are pretty common around here, but a very low-powered pump would increase their effectiveness considerably (and I would obviously also be interested).

    I have seen advertised a Peltier-device powered fan for increasing the convective heat-transfer from a wood-stove. This is cool:

    Thoughts? Leads?

  8. Netcrawl
    June 25, 2014

    @matthias I agree with you about this difficulty, there's some areas that need to be addressed- materials, electronic, communication technology, techniques and device construction. We need to deal with some serious challenges such as how to incorporate energy harvesting functionalilty into compliant flexible materials with minimal impact to wearer, how to find the best location in the human body.    

  9. geek
    June 25, 2014

    “But, engineer needs to determine how much energy the wireless sensor requires averagely as well as in downtime before developing energy source”

    @DaeJ: I agree with you that the engineers also need to figure out what happens during down time and what other form of energy will be available in case the main sources (in this case mechanical vibrations) are not there. I think that's an equally important aspect.

  10. geek
    June 25, 2014

    “Can such temperature difference based harvesting be used in medical device implants”

    @Jayaraman: I think that sounds practical, at least theoretically. What I'd be more concerned about is the relaiability aspect. Even if the device is being powered through internal movements, there has to be a back up source of power that can kick in if something goes wrong with the primary source. For instance, you can have a battery which charged constantly through the vibrations.

  11. geek
    June 25, 2014

    “From my perspective, this looks like a relatively simple and high ROI project; I am surprized I can't just go out and buy it! The regulatory/safety issues around messing with the gas-supply are all that worries me against building my own.”

    @GSKraise: It seems like an interesting idea to me. Do you have any idea about the initial amount of investment in producing this technology commercially? Also, who are some of the companies who may be looking to invest in the project?

  12. GSKrasle
    June 25, 2014


    I haven't actually looked a whole lot into manufacturers, just questioned distributors, who have never heard of equipment like I'm wishing-for, and looked at a few catalogues.  

    If I got a product already equipped with a millivolt system, I could certainly add the capability to regulate the tank temperature at any convenient point lower than its native thermostat, but the necessity of the continuous pilot (heating the thermopile) would be hard to get around. It would be a nontrivial exercise to design a way to run the pilot valve, main valve and some sort of pilot igniter from the thermal difference between the tank and ambient, especially if the tank had cooled a lot or had been run at reduced temperature to save energy. For safety, the design would also necessarily incorporate a flame/heat detector to determine when/if the pilot became ignited: releasing gas continuously without flame would be dangerous (especilly since pilots in millivolt systems are apparently large in order to drive the thermopile).

    If I got a product already equipped with electrical ignition, it would presumably have a connexion with the AC mains, and little incentive for electrical efficiency: the igniter and valves might be out-of-reach as far as using heat-harvesting to power them. That brings-up a big question I have had for a long time: why are the modern high-efficienct hydronic heating and hot-water systems NOT designed to be functional during power-outages? Is everybody expected to install back-up generators? My stove (but not its piezo igniters) works fine without electricity… but it's very old.


  13. RedDerek
    June 26, 2014

    Years ago a buddy and I thought about this product and the efficiencies were not good enough. Now for those outdoors people, this is the hot new product.

    It was a project a couple of years ago and has taken off quite a bit with others jumping in. Now I am just waiting for a good price.

  14. yalanand
    June 30, 2014

    I believe that they should come up with devices that are safe and that will guarantee that there will be no harm to the user. They should also think of coming up with a backup source that will serve as security in the case of a power failure. Important issues that need to be addressed concern the materials and the device construction. The device is expected to be constructed in a certain way to ensure that it is safe and secure. If they are planning on harvesting any energy they should ensure that it is thermal delta.

  15. SunitaT
    June 30, 2014

    @RedDerek, Powerpot seems to be good product. There will be lot of such products, which will make our life better, will be availale in future.

  16. vasanjk
    July 15, 2014


    Agreed. In order to ensure continuous seamless functioning, such implants work on a reliable energy source packed with necessary power. Thermal differences inside the body may be very minimal and may not be enough to generate necessary voltages let alone the current and hence the power.

    You guys are into something which could change the life of b patients forever.





  17. vasanjk
    July 15, 2014


    I am just thinking widely- Can the battery be charged by swinging the arms or legs,whichever is convenient to the patient?

    The calories spent could transform into the energy needed to charge the battery- I am yet to figure out how but as I said it is wild thinking.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.