Christian Promper, alumnus of the master course Information Technology & Systems Management, was awarded with the “AK Wissenschaftspreis 2017” for his thesis titled “Anomaly Detection in Smart Grids with Imbalanced Data Methods” supervised by Dominik Engel.
This year’s research question of the Salzburg Chamber of Labor was how to improve the working and living conditions of workers in terms of distributive justice.
Detection of anomalies in Smart Grids
In his thesis, Christian Promper dealt with anomaly detection in Smart Grids. The detection of anomalies plays an important role for the trouble-free future energy supply with Smart Grids.
“Currently there is little experience in smart grid anomaly detection. Since anomalies are rare, the use of common detection methods causes poor detection of anomalies, because the imbalance results in significantly more data items being attributed to common behaviour rather than anomalies within smart grids,” says Promper.
As part of his work, he therefore examined various methods to improve the detection rate for rare irregularities in data sets. In a first step Promper examined various ways to account for imbalanced data in general. In addition, Promper created a three-layer smart grid architecture with intrusion detection systems using imbalanced data methods at each level. This proposed approach outperformed common methods.