Continuous and Non-Intrusive Energy Monitoring Challenge: Smart Choices in Difficult Situations

This paper introduces the challenges tackled by power quality monitoring systems in their process of adaptation to the modern power grid (smart grid). The focus is placed on the need for new tools for the visualization and analysis of the electricity supply in a more intelligible way and from the prosumer's perspective. The working hypothesis is the continuous monitoring of the voltage signal for the extraction of statistical features. The time domain can be approached with greater computational ease and at the same time reduce the complexity of the extracted information, introducing other measurement indicators. This is achieved by employing data visualization techniques that allow for a greater understanding of the end user/system. Solutions that incorpore higher-order statistics (HOS) into the analysis helps to characterize the signal more than the second order of analysis. The work presents three sets of results. First, specific tools for management consultation with a database. Secondly, an automatic classifier of three types of electrical disturbances, based on unsupervised learning using K-Means, exhibiting a 95% detection rate. Finally, a network status monitoring system based on a nanogrid, that implements 2D graphics with HOS as a measure of supply reliability zones accompanied by a new global benchmark index. The results show three power quality concerns, whose challenges include adapting to current regulations.