NTL Detection in Smart Grids by means of a Reservoir Computing-based Solution

Smart grids are ushering in a transformative era for energy distribution and consumption, yet their emergence also brings forth novel security and fraud detection challenges. The intricacy of detecting fraud within smart grids demands sophisticated techniques for scrutinizing vast volumes of time series data. This paper introduces a novel approach that amalgamates time series aggregation functions, time series clustering using the Spearman's distance, and reservoir computing forecasting to effectively uncover fraud within smart grid systems. This is then compared with the real values to classify each prosumer behaviour as regular or fraudulent. The approach is validated by means of data collected from the Parc Bit distribution grid, located in the outskirts of Palma (Balearic Islands), Spain. The results obtained, including the comparison with previous works, demonstrate the effectiveness of the approach proposed, shedding light on its promising potential. In more detail, we show its ability to reduce the false positive rate while maintaining a high true positive ratio, resulting in an increased AUC score. As a net effect this helps to mitigate financial losses and address the various impacts associated with fraudulent activity on smart grids.