Pisani, Cosimo (2014) Real Time tracking of electromechanical oscillations in ENTSO-e Continental European Synchronous Area. [Tesi di dottorato]

[img]
Preview
Text
pisani_cosimo_26.pdf

Download (4MB) | Preview
[error in script] [error in script]
Item Type: Tesi di dottorato
Lingua: English
Title: Real Time tracking of electromechanical oscillations in ENTSO-e Continental European Synchronous Area
Creators:
CreatorsEmail
Pisani, Cosimocosimopisani@gmail.com
Date: 25 March 2014
Number of Pages: 172
Institution: Università degli Studi di Napoli Federico II
Department: Ingegneria Elettrica e delle Tecnologie dell'Informazione
Scuola di dottorato: Ingegneria industriale
Dottorato: Ingegneria elettrica
Ciclo di dottorato: 26
Coordinatore del Corso di dottorato:
nomeemail
Serpico, Claudioserpico@unina.it
Tutor:
nomeemail
Lauria, DavideUNSPECIFIED
Domenico, VillacciUNSPECIFIED
Giorgio Maria, GiannuzziUNSPECIFIED
Date: 25 March 2014
Number of Pages: 172
Uncontrolled Keywords: Electromechanical oscillations; ENTSO-e; Hilbert transform; WAMS.
Settori scientifico-disciplinari del MIUR: Area 09 - Ingegneria industriale e dell'informazione > ING-IND/33 - Sistemi elettrici per l'energia
Aree tematiche (7° programma Quadro): ENERGIA > Reti di energia intelligenti
Date Deposited: 13 Apr 2014 20:51
Last Modified: 27 Jan 2015 10:37
URI: http://www.fedoa.unina.it/id/eprint/9665

Abstract

Small signal stability is a crucial aspect to accurately keep under control in modern interconnected power systems in order to ensure their security and reliability. Such an aspect could represent a serious limiting factor in the search for ever higher power systems exploitation levels. Power oscillations not well-damped may jeopardize the system integrity on large scale: several incidents caused by the establishment of large oscillations have been recorded in the past around the world. Therefore, a basic assessment that must be done before setting a certain optimal operational framework is the determination of the actual dynamic stability margins. The fast deployment of measurement and instrumentation facilities provided by the Wide Area Measurement Systems (WAMS) technology offers a valid support in this sense. Large amount of data coming from Phasor Measurement Units (PMU) installed in the key points of power systems (e.g. primary substations) increases the Transmission System Operators (TSO) situational awareness. Thanks to accurate and timely information the stability margins can be precisely determined and optimized so that power systems can be operated at their actual full capacity while staying within the stability boundaries. A deep investigation about the WAMS currently in operation or under testing around the world confirms how power oscillations tracking is one of the main functionality/application envisaged in these architectures. Real time detection of dangerous power oscillations and hence their related continuous parameters estimation, in wide area sense, is vital in the framework pointed out above. The output of this task is therefore represented by estimates of the oscillations fundamental parameters (e.g. frequency, damping factor/ratio, amplitude and phase). If potential unstable phenomena are detected (e.g. estimating a damping ratio lower than a certain threshold value) all the necessary countermeasures have to be implemented for restoring secure and stable operating conditions (e.g. generators’ re-dispatch, tie line flows adjustment, load reduction, network topology change etc.). It was moreover found that the major problems which characterize these infrastructures rely on their own technological complexity, on the data management but especially on the research of robust identification techniques for implementing all the Dynamic Security Assessment (DSA) tasks that must be run in parallel in the central control centres. In this regard, two fundamental approaches could be applied for tracking the electromechanical modes in an electrical power system. Model-based methods (a.k.a. Component-based method), which use an electric power system model linearized around a certain equilibrium point to identify the electromechanical modes characteristics through eigenvalue analysis (whose chief rudiments are reported in the Chapter 3). Eigenvalue analysis is not suitable for on-line tracking, especially for large scale power systems due to both high computational time and uncertainties in power system modeling. Measurement-based methods (a.k.a. Mode Meters), estimate an updated model of the electric power system from direct system measurements which come from measurement devices installed on power systems. These techniques, freeing themselves from the system modeling, they consider the power system as a black box and by making use of the signal processing expertise, estimate the modal content of the acquired signals. Being moreover less expensive than the first class of methods in large scale power systems model set up, it appears clear that they are suitable for an on-line DSA task. However, the set of available measurement-based estimation techniques is fairly wide. Besides I note that relevant journal databases are regularly filled by novel more and more advanced algorithms. My personal feeling in this regard is that the basic methodologies are really few, while several refinements of the same algorithms, aimed at overcoming specific weaknesses, are regularly proposed. From the experience gained working hardly on the topic I can state that no best estimator exists due to the lack of an accepted definition of optimality. Furthermore, it is a difficult task to assess the performance of different estimation methods because each of them was initially designed for a specific field, has its own features and sometimes presents parameters chosen according to experience or through heuristic considerations. This means that for instance a method could show good performance in damping and frequency estimation if the modes number is known while may fail if it is not know in advance. In addition, a method could work better than another for noiseless sampled signals while could deteriorate its efficiency when the signal-to-noise ratio (SNR) decreases. Nonetheless, there exist estimation techniques which are “generally” characterized by good performance with respect to the others. The meaning of the term “generally” should be intended as “with respect to the main situations that may occur”(different data typologies, various SNR levels, a priori knowledge of the intrinsic power system modes etc.). A wide set of estimation techniques will be analyzed in the present thesis. Afterwards, a performance comparison among the techniques will be accomplished with the objective of pointing out strengths and drawbacks of each of them. Once ascertained the points to improve, three novel estimation algorithms will be introduced. They represent a good complementary tool to the ordinary model-based methods implemented in the central control centres for real time monitoring power system oscillations. Almost all the estimation algorithms considered in the thesis were tested in real time on the Italian WAMS thanks to the support of the TSO, Terna. The complex infrastructure owned by Terna, thanks also to the real time information exchange with some European partners, represents a vigilant eye on the entire European Network of Transmission System Operators for electricity-Continental European Synchronous Area (ENTSO-e CESA) for the purposes of analysis. The emphasis of this research was hence to tailor high accurate and resilient estimation algorithms for real time monitoring of electromechanical oscillations, in particular of inter-area type, in such a large interconnected system. Although the doctorate course ends achieving the predetermined objectives the research on the topic will continue.

Actions (login required)

View Item View Item