Load profiling with fuzzy self-organizing algorithms

This paper describes an enhanced fuzzy self-organizing algorithm (EF-SOM) to address the problem of consumer classification in electric distribution networks based on the shape of the load profiles (LPs). This algorithm is a modified form of the standard fuzzy Kohonen algorithm, which determines the deviations between metered-LPs and the prototypes of the typical LPs using weighted windows around the peak and valley hours. The EF-SOM algorithm was tested on an independent load profile database, proofing its ability to filter outliers LPs and to produce realistic classification results.