The aim of NRG4Cast project is to develop advanced solutions for predicting behaviour of local energy networks for the three fundamental scenarios:

  • Predicting energy demand on several network granularity levels (region, municipality, city, business, household and energy service provider),
  • Predicting energy network failures on interlinked local network topologies,
  • Detecting short term trends in energy prices and long term trends in national and local energy policies.

The main concept is monitoring of energy behaviour in different network layers; from single buildings, to building block, neighbourhoods and urban environments.

(Local network topologies for energy demand behaviour analysis)

The system will collect and monitor a holistic view of energy consumption information from the micro-device level consumption in a local context to an aggregate view of energy grid level usage. Different types of users of the system will have different objectives:

  • Local Governments may use the system to optimally plan, negotiate and buy energy based on the predicted local demand and energy prices,
  • A Distribution System Operator may use the system to plan contracting with neighboring distributors to optimize seasonal, peak or event driven demand,
  • Governmental agencies/energy policy makers may use the system to monitor energy consumption and demand forecasting, that will help in shaping the energy policy in the specific region (e.g. new investments in energy sources),
  • A local utilities company many use the system to offer flexible pricing based on supply and demand, events and incentives,
  • A building generating excess power through renewable sources may choose to use the system to determine the appropriate moment to sell excess energy to the grid or store in local storage,
  • Citizens may refer to the system to plan home usage of high consumption appliances, and charges for of electrical vehicles, automobiles, etc.,
  • Smart devices or energy home gateways may consult NRG4Cast services before selecting the best energy supplier for a particular scenario.

To realize the above operational model, NRG4Cast will develop and deploy advanced technologies from the wide area of Artificial Intelligence research, ranging from machine learning, data (SVMs, GMDHs, etc.), text and stream mining, complex event detection methods, prediction methods (from statistical time series to naive and probabilistic methods). All methods will have to act in real-time with mass amount of data that will be available in multiple forms (data from sensors, extracted information from public sources, structured knowledge).