[IOT] Smart meter

Inspiration

We love connected devices and measurement.

What it does

This device measures power consumption of household appliances, does analysis of the results, stores them in database and applies machine learning to identify the appliances. This can help to make more detailed power consumption and help power distributors to react accordingly.

How we built it

We built in three layers. Lowest is smart meter at every outlet, sampling data, calculating power and analysing current haromnics. Above power meters is aggregator, linked to meters via ZigBee, gathering measured and preprocessed data; seding to server. Server does the machine learning appliances sorting, logging to database and presenting user friendly data via web UI.

Challenges we ran into

Wi-Fi was quite slow, we had to set up our own AP. We had to select ZigBee channel to avoid Wi-Fi traffic.

Accomplishments that we're proud of

It works.

What we learned

We learned to work wit ellastic search, FLASK, networking.

What's next for Smart meter

Implement more security. Implement power consumption prediction - currently we are doing long term prediction over existing data from http://redd.csail.mit.edu/ Implement more sensors

Built With

• arm
• c++
• elasticsearch
• flask
• kibana
• linux
• pandas
• python
• raspberry-pi
• sci-kit-learn
• zigbee