Local and Wide-Area Sliding Mode State Observation, Fault Reconstruction and Control with Application to Modern Power Systems

The present thesis summarises the research activity carried our by the PhD candidate Gianmario Rinaldi from October 2016 to September 2019 at the Department of Electrical, Computer and Biomedical Engineering at the University of Pavia, Italy. In recent years, radical changes are taking place to power systems. A worldwide consensus has been reached for the reduction of greenhouse effects, by promoting the growth of renewable energy sources in power grids. Therefore, a special shared interest has been raised amongst researchers and practitioners to turn the existing power grids into smarter and more reliable ones, which are able to safely, efficiently, an reliably integrate the growing renewable energy sources. Supervisory Control And Data Acquisition (SCADA) has been the conventional control and state estimation methodology practically used in the last few decades. Recent progress in computer science and electronic technologies has opened the door to the implementation of the so-called Wide Area Measurement Systems (WAMS). In particular, with a widespread deployment of Phasor Measurement Units (PMUs), a more accurate depiction of the state in power systems has become achievable in practice. Latest advances in computer science and electronic technologies have laid the groundwork for the conception of the so-called cyber-attacks, which can be defined as computer-based algorithms capable of destabilising the power network by compromising the collected measurements to be sent to a control centre, attack the communication networks, or alter and delay the control variables. In order to turn the existing power system into a smarter one capable of both harmoniously integrating renewable power sources and efficiently and safely dealing with faults and cyber-attacks, the attention is now focused on the following relevant research areas: The design and assessment of more accurate, robust and dynamic state estimators in power systems, which are able to obtain a near-real time depiction of all the state variables, instrumental in enhancing the monitoring of the networks. The implementation of timely fault detection, reconstruction and mitigation methodologies, devoted to preserve the stability of the entire power network, thus preventing wide-spread outages, blackouts, and degradations of the power quality. The design of identification schemes to determine key-properties of the components in power system context. The design and assessment of novel control approaches devoted to both regulate the frequency deviations and minimise the cost of the power generation. These control schemes are also required to be robust to possible faults, disturbances, and uncertainties affecting the power systems. The present thesis aims to fit into the aforementioned promising research areas in power systems. In particular, four challenges are addressed: The first addressed challenge considers the design of robust state estimators which are able to depict in near real time the state of the overall power systems to globally enhance the monitoring of the power systems, thus reducing the number of the deployed sensors. The undertaken analysis started at the local level and then consider the power system as a large-scale system. The second addressed challenge focuses on the design of fault detection, reconstruction, and mitigation approaches devoted to enhance the resilience of the power network. The third addressed challenge considers the design and the assessment of robust sliding mode observer-based controllers which are capable of regulating the frequencies in power systems whilst minimising the cost associated with the generators. Finally, the fourth addressed challenges examines the identification of the relative degree properties with application to electrical and power systems frameworks. Furthermore, the outline of the present thesis is coherent with the development of the contributions illustrated above.