TRY TO PREDICT GRID FAULTS: A DYNAMIC, SEMANTIC-BASED AND MULTI- DIMENSIONAL APPROACH

The analysis of big data volumes, produced by energy distributors and concerning the operation of the grid, requires several different techniques, particularly if one would like to use such large amount of information in possible grid failure assessment and even prevention. A preliminary data volumes reduction is strongly needed: currently grid data need huge storage and processing resources for long-term analysis. This paper presents a data aggregation statistical framework allowing to compact grid data sets in order to enable further semantic-based decision support (grounded on logicoriented reasoning technologies) and devoted to try to predict failures, to suggest targeted maintenance and to possibly optimize grid assets.