Field-proven machine learning and artificial intelligence engine provide an automatically tailored forecasting model for each building and asset to create bottom-up predictions on a minutely basis.
A unique data recovery system allows to support both smart meters and legacy meters, and Big data infrastructure provides scalability for up to 1 sec data interval.
Heuristic tailor-made algorithm for optimal strategy search takes into account price fluctuations, weather conditions, and available consumption flexibility. The system employs probabilistic techniques to handle different possible scenarios of nano-grid, micro-grid and VPP-level energy balance.
The optimization engine supports multi goal optimization of cost, emissions, and convenience as well as planning tools for most efficient DER deployment.
To achieve high performance, scalability, and resiliency of the system, Energy A.I. is built in a hybrid Fog architecture. This approach uses a combination of centralized computing capability for forecasting grid and prosumer behavior together with decentralized and autonomous agents for local optimization.