Modern power distribution systems have migrated from a passive, non-supervised structure with low level of automation to an active, supervised structure with high level of automation due to the introduction of photovoltaic (PV) generators, dispersed smart sensors/meters, and new automatic control devices. Despite the technical and environmental advantages, these developments may create new adverse concerns on distribution systems. Examples of new concerns are resonances and instabilities between capacitors connected in low voltage (LV) systems and PV generators; or control conflicts between multiple devices performing voltage control actions such as voltage regulators, switched capacitors, PV inverters etc. Simultaneously, the massive amount of data (Big Data) provided by the modern measurement infrastructure contains valuable information to anticipate, detect, and mitigate these newfound issues. However, actionable information is not readily available in the raw data. In this context, the main goal of this Ph.D. project is to develop methods for system and equipment condition monitoring based on data analytics techniques by using data provided by the measurement infrastructure available in the modern distribution systems. Four specific issues will be investigated: resonances and instabilities in LV systems, resonances and instabilities in medium voltage (MV) systems, voltage control conflicts in LV systems, and voltage control conflicts in MV systems. For each of these problems, the objective is to use measurement data to (a) anticipate the risk of a given problem, (b) detect the problem in progress quickly, (c) determine the extent of its impact in the system, and (d) develop mitigation strategies to avoid it. Extensive simulation studies and field measurements will be used to validate the proposed techniques. Finally, the minimum measurement requirements to apply each proposed method will also be established. The development of these condition monitoring techniques and full characterization of the new system problems will increase considerably the situational awareness of the Distribution System Operators (DSOs) so that predictive mitigation actions can be adopted as soon as a risk situation is detected.
News published in Agência FAPESP Newsletter about the scholarship: