Tessitore, Salvatore (2022) Detection and Measurement of inter-area oscillations for power system stability. [Tesi di dottorato]

Anteprima

Testo Tessitore_Salvatore_XXXV.pdf Download (11MB) | Anteprima

Abstract

Interconnected electrical system stability is the ability of the system to find a new equilibrium condition following a disturbance. The phenomena of stability of the electrical system can be classified into three categories: rotor angle stability, frequency stability and voltage stability. In this doctoral research activity, the focus is on low-frequency oscillations (LFO), which are phenomena related to the stability of the rotor angle. This is a crucial aspect to be carefully monitored in the Italian and European electricity system, to ensure its safety and reliability. Undamped frequency oscillations can compromise the system integrity on a large scale: in the past, several accidents have been recorded caused by the establishment of high-intensity oscillations around the world. Thanks to new technologies, in recent years measurement and instrumentation structures based on Wide Area Measurement Systems (WAMS) technology have become widespread, which is essential to monitor and characterize this type of phenomenology. The possibility of synchronous and other frequency sampling, the possibility of a communication protocol capable of transmitting high sampling data with low latencies guarantees a large amount of data from the phasor units of measurement (PMUs) installed throughout the European electricity system, increasing its observability by transmission system operators (TSOs). A description of the WAMS currently used by the Italian TSO (Terna) confirms that the detection of frequency oscillations is one of the main functionalities/applications planned by these architectures. The real-time detection of dangerous frequency oscillations and then the estimation of the relative parameters (frequency, damping, amplitude and phase) is fundamental in the framework described above. When potential divergent oscillations are detected, all necessary countermeasures must be implemented to restore safe and stable operating conditions (e.g., re-dispatching of generators, regulation of connecting line flows, load reduction, modification of network topology, etc.). For this purpose, the major problems concern the complexity in the search for robust identification techniques and accurate characterizations of frequency oscillations. In this regard, several fundamental approaches for tracking electromechanical modes in an electrical system are reported in the literature. Some approaches use a linearized electrical system model around a certain equilibrium point to identify the characteristics of electromechanical modes through eigenvalue analysis. Others are based on the estimated measurements of an updated model of the electricity system from direct measurements of the system, coming from measuring devices installed on the electricity grids. From the experience gained working for a long time on the subject, I can now say that there is no optimal algorithm applicable in all operating conditions but rather each one has advantages. This means that, for example, one method might show good performance in phase and frequency estimation, another might show very good performance in estimating damping and amplitude. In addition, one method might work better than another for signals sampled without noise, while it could worsen its efficiency when the signal-to-noise ratio (SNR) decreases. However, there are estimation techniques that are generally characterized by good performance compared to others. In the present thesis, first different estimation techniques have been analyzed both on simulated data and on real data. Subsequently, improvement solutions have been proposed compared to that reported in the literature and finally in relation to the monitoring or defense objective set with the Italian TSO, the most appropriate method for real-time application has been chosen. The goal of this research was therefore to create highly accurate and resilient estimation algorithms for real-time monitoring and defense of electromechanical oscillations, particularly inter-area, in such a large interconnected system. Although the PhD course ends by achieving the established goals, my research is still ongoing as an employee of the Italian TSO (Terna).

Downloads

Actions (login required)

Modifica documento