predictions

Inspiration

A desire for open exchange of data/information to improve overall usage and adoption of solar/battery technologies.

What it does

Makes real-time decisions of optimal behavior for controlling solar/battery/grid energy.

How we built it

Using real historical data from commercial buildings, PV generation, and battery behaviors. We integrated these into a webapp built in python.

Challenges we ran into

Scoping, depth of subjects.

Accomplishments that we're proud of

Specing a robust software architecture with naive models in 8hours. Providing a flexible architecture that adapts to variety of assumptions/strategies.

What we learned

The most complicated aspect of these integrations is the Battery management. As was mentioned several times today "the science isn't there yet", so this is a huge difficulty and opportunity for machine learned improvements in behavior.