Research On Fault Component Localization Of Electric Power Systems Based On Wide Area Measurement Information

The key issue of novel wide area backup protection has been concentrated on the rapidly and accurately fault component location. The location method used in this novel backup protection system is mainly relying on the voltage and current synchro-phasors provided by the Wide Area Measurement System, WAMS. Obviously, the faulted component should be determined as soon as possible, which must not exceed the limit of time delay in design principle of backup protection. And then, it is the rest of time that is enough to adjust the setting of related protections. In this paper, the research of novel faulted component location method based on WAMS will be the focus, on which a lot attention has been put. The main contents of this thesis contain:(1) According to different types of fault in complex electric power system, a cluster analysis based fault location scheme has been provided. In this method, fuzzy cluster analysis and hierarchical cluster analysis and so on are used to classify the changes of electrical quantities from system failure. Finally, fault component location and fault section partition are realized accurately and effectively.(2) A pattern recognition based method for fault location in complex electric power system is proposed. The pattern classification technique and discriminant analysis principle in pattern recognition theory are utilized to study the inherent law of electrical quantities' marked changes from system failure. The simulation results indicate that respectively study on the phase or sequence voltage and current in unsymmetrical faults and symmetrical short circuit faults, the pattern classification technology and linear discrimination principle are able to rapidly and accurately identify the fault components and fault sections.(3) Based on principal components analysis theory, two fault feature extraction schemes are formed by real-time measurement information of PMU in complex electric power systems, namely, fault feature extraction scheme based on node coefficient in principal components and fault feature extraction scheme based on node score of principal component. According to "Scheme A:node coefficient", by analyzing node coefficient, the fault is usually corresponding to the variable with the biggest coefficient in principal component. According to "Scheme B:node score", the fault is usually related to the node with the biggest principal component score. In a summary, both of these two schemes are able to carry through fault detection independently and effectively.(4) We have thoroughly studied nonlinear complex dynamical system from the layers of theory and application. In order to quickly and accurately locate faults in complex power grids, many explorative researches have been initiated by means of symbolic dynamics. According to the symbolic dynamics analytical results of PMU real-time measurement information from WAMS, we have advanced exact and quick fault localization by computing Lyapunov exponent. The studies indicated that the actual fault position is usually near the position with the biggest Lyapunov exponent during fault circumstance in complex electric power systems. Generally, the nearer to the fault position, the bigger Lyapunov exponent is. The obtainment of this conclusion may represent an inherent and essential attribute of electric power system.