Blogs

Do we need to build nuclear power stations to cope with the IoT Energy demand? – Part Two

IoT devices are often built to strive for efficiency. If you remove the huge, power hungry screen of a mobile phone the energy consumption will naturally be lower, write Abilio Marques, a senior software engineer, and Bettina Rubek Slater, the chief operating and chief marketing officer of Pycom, in the second part of their blog.To continue with the idea, let’s introduce an example. Imagine a temperature sensor that reports its measurements every 20 minutes over the cellular network. A typical device with a 3G radio will use around 60mW (milliwatts) during standby (idle). Each packet of data transmitted will use around 12 Joules. If you are transmitting data three times an hour then you’ll need the equivalent of an additional 10mW on average.

That means that every hour the sensor will use around 70 mWh to transfer data. Powering 25 billion of those devices will use 1.75 GWh.

That’s a lot of energy, and it’s the very same reason engineers have been working on alternatives. We call those low Power WAN with networks such as LoRa, Sigfox and Narrowband IoT (NBIoT). They are all a new breed of wireless networks that have been designed from the ground up with energy efficiency as one of the main goals.

Taking our sensor again, what difference will LoRa and hardware that is designed to use as little energy as possible make? Well, the Pon (power when on) will be around 70mW while transmitting – that’s assuming three seconds each hour are used to transmit. This is comparable to the standby power on a 3G cellular network.

The magic of these IoT devices, however, is that they can use very little power while being idle. Actually, a device can use as little as: 600nW (that’s nano Watt, as in billionths of a Watt) and 60 uWh (microwatts hour, millionth of watt per hour). That’s three times less energy than using regular 3G cellular devices. Powering 25 billion of those devices will consume 1.5 MWh (megawatts hour). That’s an improvement of more than 1000 times.

So is the answer to build more nuclear stations to power the future IoT world?

It’s not for us to decide on the source of energy to help reduce the CO2 footprint.

Today, there are more than 436 commercial nuclear reactors operating in 30 countries. They produce the equivalent of 11% of the world’s electricity. In fact, it is claimed that the world now produces as much electricity from nuclear energy as it did from all sources combined in 1960. Around 19% of the UK’s energy supply is from Nuclear power. We understand that the average nuclear power plant produces more than 600 MW per reactor per year.

From a theoretical point of view, if we use the already mature 3G cellular networks, we’ll need to find a way to produce the extra 1.75 GWh every hour to power the 25 Billion new connections.

And nuclear is a potential option.

However, if we choose the more efficient technologies like LoRa where we have 1000 times more efficient use of energy, and even if we increase the frequency of the reporting to and from the device, the consumption will stay close to the calculated 1.5 MWh. These can be powered by small coin sized batteries, wind or solar and remain well bellow a small commercial reactor output, which is around 600 MWh.

That way, no extra nuclear power plants are likely to be needed.

Final thoughts

It is also important to consider the way in which we power IoT devices as keeping them running from the mains is not a logical option. Today’s most efficient wall warts will use an idle power of 100mW, even if nothing is connected to them. Compare that to even the smallest LoRa sensor with 60uW needed to power it. If you then want to keep a small sensor running, you will use 100.06 mW of which only 0.06mW is powering the sensor itself. More than 99.9% of the energy consumed is wasted in heat.

By running IoT devices of the mains, regardless of whether these are Cellular or LoRa based, at 100mW for each wall wart, we would add 2.5MWh to the current consumption further amplifying need of new power sources.

None of our calculations have taken recycling and manufacturing energy into account. But it doesn’t mean it’s not important and in time we’ll hope to see the conversation evolve from how much power is needed to run a device, to how we can combine solar energy as technologies becomes more advanced. If we can start thinking about these scenarios now as the IoT develops we’ll all be doing a better job of connecting things while protecting the planet.