Smart State Technology

Smart State Technology provide synchronised measurements and an open distributed platform to develop & execute real-time (sub-second) algorithms on the electrical power network.

They have worked and work in research projects related to matching green energy with demand, time-shifting loads and fair use of electrical connections while mitigating network congestion.

They have used machine learning to identify characteristics or patterns for certain loads (such as in-house appliances and PV inverters). They did an experiment to use machine learning on real-time raw signals (4kHz) to replicate the synchro-phasors they create with an algorithm. They have created a general purpose building block in their platform that can run a trained ML model on (combinations of) intermediate data streams from different locations (e.g. 4kHz raw data, 50Hz voltage/current/power phasor, 1Hz harmonic distortion, etc.). Their University partners use ML for demand-side management, profile steering, and load/generation forecasting.