Font Size: a A A

Study On Fault Detecting And Stable Control Of Power System Based On Information Fusion

Posted on:2006-06-09Degree:DoctorType:Dissertation
Country:ChinaCandidate:C X DouFull Text:PDF
GTID:1102360152495552Subject:Measuring and Testing Technology and Instruments
Abstract/Summary:PDF Full Text Request
This thesis researches on application of information fusion theory to power systems in order to settle some of the problems existing in the great and complex energy systems. Two aspects of research works have been done in this thesis. One is on constructing integrated fault detecting criterion based on information fusion for single phase to ground fault (SPGF) in power distribution systems to overcome the disadvantages in it as very weak signals and complex fault situations. The other aspect of research work is on exploring a decentralized H^ fuzzy tracking control scheme based on information fusion techniques to improve the robustness of stable control for multi-machine interconnected power system, which is characterized as huge dimension and strong nonlinearity.The SPGFs occurring in Neutral small current grounding (NSCG) power systems, due to the very little fault currents they produce and the intermingle with some other adverse factors, are difficult to be detected. Though lots of works on it have been done in the past, this problem has not been settled perfectly as yet. The research contents on SPGF detecting problem in this thesis are as follows:In order to take full advantage of multi independent information of SPGFs to construct integrated fault detecting criterion, Multi-criteria fusion method based on D-S evidence theory is studied, and a integrate fault detecting strategy is proposed.In order to make the best of abundant fault transient information, a neural network fault detecting method based on transient information is proposed. Firstly, a appropriate orthogonal wavelets packet function is chosen in order to effectively distill fault transient information; and then according to the characteristics of the fault transient information of power system, a chaotic neural network is investigate and employed to detect SPGFs, by which the interference of false and non-fault transient information is overcome, its weight coefficient and parameters are optimized by improved genetic algorithm; Finally, according to relation between fault aim patterns and nerve center output states, a numerical type of fault detecting criterion is designed, which ensure the rationality and probability of multi criteria fusion. Effectiveness and advantage of the proposed method is tested by several experiments.In order to make use of the independent information, a multi-frequency bands analysis fault detection method based on singularity of fault signals is developed. Firstly, an appropriate cubic B-spline wavelets pocket transformation is used in order to effectively transform the singularity of fault signal to the maximal moduli in wavelet domain. Second, the selected frequency bands for neutral-isolated system and compensated system are studied based on the analysis of angle-frequency characteristics. And then the characteristic frequency band of each line, in which the maximal moduli is most concentrated, is chosen according to the view of maximum energy. When the characteristic frequency band of each line is same, three numerical type of fault detecting criterions are designed, fault detecting principle is constituted; When the characteristic frequency band of each line is different, two numerical type of fault detecting criterions are designed, fault detecting principle is constituted too. Several experiments show that the proposed method is effective and particular.Utilizing the steady state fault signals, a chaotic theory based SPGF detecting method is proposed. The background signals are modeled with fuzzy neural network (FNN) based on chaotic theory, and through the prediction of the FNN, the background signal in fault can be filtered to remove the interference of background signal as 5th harmonics. The Duffing oscillator chaotic detecting principle is employed for extracting of very weak 5th harmonic signal, and the corresponding method as 5th harmonic phase altering is established. For sake of fusion, this detecting method is converted into numerical type of fault detecting criterions. The experiments show that the proposed method can choose out the fault line exactly and reliably.As the emerging of large power system inter-connecting and the wide area application of power electronic devices to power systems, the complicacy of power systems have increased greatly. Stability control for multi-machine coupling power system has exhibited a critical and formidable problem. Considering a nonlinear, giant dimensional and strong coupling power system can be equalized as interaction of several single-machine subsystems, therefore the stability control of multi-machine power systems can be implemented through the decentralized coordinating control to the subsystems in view of the interaction among them. Each coupled subsystem operates in different status, and behaves different features. Therefore, the control schemes to the different subsystems should differ; otherwise, they will depress the robustness and even the stability. This thesis proposes a...
Keywords/Search Tags:information fusion, power systems, evidence theory, wavelet theory, neural network, Duffing oscillator, fuzzy tracking control, fuzzy observer
PDF Full Text Request
Related items