Detecting Unknown Electrical Loads Using Open Set Recognition

This paper presents a solution to the problem of detecting unknown electric loads in a household. With the emergence of renewable energy sources, the balancing of electricity production and demand has become more challenging, and control of the demand-side of the grid is needed. This requires knowledge about the load composition and control of the loads. For this task, knowledge about the loads is required. The problem with current load classification solutions is the closeness of the classification. Most smart plugs on the market cannot detect the loads. Moreover, the load classification solutions available in the literature lack the ability to detect if a previously unseen load is connected. We present a novel solution to this problem by applying an SVM-based Open Set recognition method to a load classification solution with already demonstrated results.