George Hébrail - Smart grid data management and mining

The talk will be divided into three parts :

- (1) a short recall of the definition of a smart grid and the different types of data processed in this context ; 
- (2) the main problems in data management in a smart grid ; 
- (3) the main problems in data mining processing in a smart grid.

For each of these three parts, the focus will be on what is new compared to legacy systems in terms of types of data, on whether the needs lead to batch or real time processing, and on whether data mining algorithms may be centralized or distributed. This presentation will be illustrated by some typical problems associated with smart grids, like monitoring the grid or forecasting of electric power consumption or production.