Dynamic forecasting and adaptation for demand optimization in the Smart Grid

The daily peaks and valleys in energy demand create inefficiencies and expense in the operation of the electricity grid. Valley periods force utilities to curtail renewable energy sources such as wind as their unpredictable nature makes it difficult to maintain line frequency across the network within target bounds. Peak periods require additional generators that remain dormant during other periods. Smoothing this demand cycle is one of the fundamental challenges of the Smart Grid, requiring flexibility and coordination between actors throughout the Grid. This paper describes the Smart Grid as a multi-layered system and proposes a cross-layered dynamic adaptation approach to facilitate this flexibility and coordination. This method uses a hierarchical taxonomy to identify appropriate adaptation actions in response to identified mismatches, supported by a run-time predictive statistical framework to predict mismatches, enabling timely adaptations to be triggered.