Data Analytics and Machine Learning for the Operation and Planning of Distribution Grids

Abstract

The recent aspirations for a more sustainable energy system and a reduction of energy-related CO2 emissions have triggered a change of paradigm in power distribution grids, often encouraged by national and supranational policies. Traditionally considered as a passive black-box component of power systems, the distribution grid currently undergoes a rapid transformation and sees the emergence of new types of loads (e.g., electric vehicles, electric heating systems, electric water heaters) as well as distributed energy resources (e.g., small wind turbines, solar photovoltaic systems, battery energy storage systems). Their integration requires increased reliability, efficiency, and adaptability of distribution systems, which inevitably relies on more visibility. Consequently, advanced electricity sensor elements are massively rolled out in distribution grids down to the end-users. The gains in transparency and controllability offered by the advanced metering infrastructure open up an extensive range of new opportunities discussed extensively in the literature. Nevertheless, the research community is usually not granted access to real-world data due to understandable privacy concerns. They must depend on simplifications and synthetic data that often do not reflect the more complex reality and might lead to biased conclusions. On the sole basis of real-world data, this thesis intends to highlight which are the assumptions that can realistically be taken in the development and validation of data-based studies and applications. It also suggests various processes and methods to effectively leverage the actual potential of the advanced metering infrastructure and address some of the current challenges in grid operation and planning. This work primarily focuses on the low-voltage level, which is still rarely considered in the state-of-the-art literature. Data preparation, big data visualization, pseudo-measurement synthesis, distribution system state estimation, load disaggregation, and short-term forecasting are among the investigated topics. In that respect, the thesis hopes to bridge some of the gaps between the relatively conservative practices in the power industry and the various advanced data-based applications proposed in the literature.