We are happy to announce that the improved Smappee NILM 2.0 was released today. This means two things: (1) some appliances in your house will automatically be labelled by Smappee itself, and (2) even some of the smaller appliances will now also be detected.
The NILM technique
NILM stands for ‘Non-Intrusive Load Monitoring’: a system that monitors the changes in voltage and current, recognises these changes as patterns and assigns these patterns to appliances. This is what allows Smappee to detect and recognise appliances in your home. The technique was invented back in the 1980s, but it was so complex that it took a couple of decades before it really got to market. And even today, it remains limited to a handful of applications in the consumer market.
For us, NILM is a groundbreaking and revolutionary technology because it allows us to give you real-time energy data down to appliance level and at an affordable price. Studies have shown that it is precisely what is needed to change your energy behaviour. In order for a consumer to initiate savings, the data provided by an energy monitor has to be real-time, down to appliance level and be actionable. For example, you will only understand that your fridge is an energy guzzler if you see in real time how much the fridge is costing you and only then will you replace it with a more efficient one.
Thanks to NILM, you need just one central device to monitor your household and all the different appliances. I mean, be honest: would you really buy and install 50 different energy meters in your home? We’re talking time and investment here.
Ever-improving appliance recognition
Smappee has been working passionately on its NILM since day one (over three years ago now), aiming at an ever-improving appliance recognition. We are now making a big step forward, gifting your Smappee home energy monitor with an updated NILM 2.0 technique to detect appliances more accurately.
First of all, Smappee can now autolabel some of the most important appliances in your house. For example, not only will Smappee rapidly detect your fridge, it will also autolabel it as ‘Fridge’ instead of ‘Appliance 1’. And that’s just the beginning; the list of appliances that can be autolabelled will continue growing in the future!
Secondly, NILM 2.0 reduces the confusion between small appliances. An improved algorithm cleans the noise generated by the many small appliances in your home, and as such, your Smappee monitor will extract cleaner patterns from the signals it hears. In real terms: an induction stove will be more easily differentiated from coffee machines or kettles.
This is just the beginning
Because it is a rather new technology, it still struggles with some rookie issues. For example, at the moment, it’s impossible to detect each and every appliance. The way NILM works can be compared to recognizing the different instruments being played at a concert: loud, prominent instruments are the first ones to be recognized. Even our NILM 2.0 still struggles to detect very small appliances (using less than 20 Watts), appliances with similar patterns (such as devices with heating elements, like a toaster, stove, kettle etc.) and appliances with an irregular consumption pattern (such as a computer or television). Nevertheless, NILM is extraordinary because it can already do so much. Our ultimate goal is to accurately detect all relevant appliances in your home. So this is a big step forward but still only the beginning.
Note. Current Smappee owners can choose to keep working with NILM 1.0 or to start over with NILM 2.0. If your Smappee monitor has been running for a while and most of your appliances have already been labelled, there is no need to restart with NILM 2.0. If you choose to restart with NILM 2.0, the total values will be saved, but you will lose the historical data of your individual appliances.
Note. To install NILM 2.0, go to 'Settings' > 'Expert Mode' > 'Appliance detection: Version 2'
Do you have interesting ideas you'd like to share with us? Mail us your suggestions, hot topics or burning questions: firstname.lastname@example.org