Studies On Some Key Problems Of Wide-area Measurements Based Power System On-Line Security Analysis

The appearance of synchronized phasor measurement based wide-area measurement system (WAMS) which is one of the new front subjects in power system can make the distributed synchronized phasor measurement, dynamic security monitoring, online security analysis and wide-area stability control possible for large interconnected power grid.Firstly, the concept of wide-area measurement system and applications of the wide-area measurement system in power system are introduced in this dissertation. Secondly, this paper has studied the wide-area measurement based real-time dynamic monitoring/state estimation, load modeling and dynamic security analysis methods. At last, this dissertation summarizes all the work which has been done in this dissertation and also proposes the potential prospect of new research directions of wide-area measurement system in power system. The dissertation consists of seven main chapters as follows:The first chapter analyzes the shortcomings of the traditional information acquisition system and traditional dynamic security analysis methods, points out the importance of the credible information, presents two concepts of wide-area measurement system and introduces various researches and applications of wide-area measurement system in power system. At end of this part, the concepts and corresponding introductions of wide-area protection and strategy power infrastructure defense system are also given out.The second chapter analyzes the optimal placement of phasor measurement unit (PMU). Due to the factors of the price and technology, now it is impossible to place PMU at all the buses in the power system. How to use the minimal number of PMU to obtain the maximal observability to power grid is the chief problem need to solve in the current wide-area measurement system. In this part, the dissertation has proposed two optimal placement methods which can meet the requirements of observablity of power grid or can catch the transient performance of all the generations, respectively. The third chapter has proposed a hybrid measurements based dynamic state estimation method. Dynamic monitoring is the basis of online security analysis, but the traditional supervisory control and data acquisition (SCADA) system can not supervise the power system in real time. The appearance of PMU makes the dynamic monitoring become possible. But because the placement of PMU of current stage may not guarantee the power grid to be observed, the hybrid measurements based dynamic supervisory or dynamic state estimation becomes the chief difficulty of current research. This dissertation uses the relationship of voltage sensitivities of buses'to solve this difficulty.The fourth chapter has proposed a wide-area measurement based load modeling method by local measurements and pseudo-measurements of PMUs'. By supposing the generator models and the transmission line models are relatively accurate, the proposed method can obtain the all load models of every load bus by using the time domain simulation and evolutionary algorithm.The fifth chapter has proposed two wide-area measurement based transient stability analysis methods. In this chapter, a stability classifier and a prediction model are set up by using of data mining technology, support vector machine (SVM) and support vector regression (SVR) models. The former uses rough set theory and self-organizing feature map (SOFM) to extract the features of the inputs and pre-classify the input samples, respectively, and at last the SVM can be used to find the optimal separating hyper-plane for the fuzzy samples, which can improve the precision of the classification. The latter can also get high prediction precision by using SOFM to classify the samples and SVR to train the samples classified.The sixth chapter has proposed a wide-area measurements based voltage stability analysis method. The local measurements can be used to assess the voltage stability problem. By using local measurements one local load branch can be seen as a simple load infinite system, so we can study the local voltage stability index at first, and then obtain the voltage stability index of the all system by synthesizing. This method can meet the requirement of real-time analysis. But the traditional equivalent method can not consider the effect of the branch capacitance and also not consider the change of load components. In view to the two defects, this dissertation ameliorates this local equivalent system based voltage stability analysis method.The seventh chapter has proposed a wide-area measurements based low-frequency oscillation analysis method. Prony algorithm is excellent method to identify the low-frequency modes, but it is sensitive to the noise. In this chapter, the dissertation compares several filter methods and proposes a new noise filter——fuzzy logic filter, analytical results indicate that the fuzzy filter is better than other methods in noise cancellation.