| Traditionally, dynamic behavior analysis of the complex power system is mainly based on the simulation study and linearized approach with detailed system model, but their accuracy depends on the detailed model and parameters. Considering the complexity and variety of the actual operating condition, the detailed model is hard to update with the change of the actual operating condition. Therefore, it may lead to a much difference between the analysis results based on detailed model and the actual operating condition. Furthermore, it also can directly affect the accuracy of the analysis and calculation of dynamic characteristics. Fortunately, the real measurement data measured by phasor measurement units (PMUs) in the wide-area measurement systems (WAMS) can reflect the actual dynamics. Therefore, it is significant to construct the mathematical model for stability analysis and control through the system identification technology, which can indicate the actual system behaviors.All the research in this thesis are based on the measurement data, a dynamic response model is constructed among measured locations to solve the dynamic response missing problem; for the inter-area oscillation, a system model is constructed and a wide-area damping controller are designed to realize the online adaptive control in the hardware test-bed. The thesis contains following several aspects:(1) In this thesis, most of the studies utilize the linearized method. Therefore, the linear characteristics are studied utilizing the linear theory at first. Based on the measurement data of the dynamic response after disturbancs, the linearity of the U.S. estern interconnection power grid and the North-Central China power grid is studied. Quantitative conclusions are obtained.(2) Dynamic response estimation models are constructed among different locations with the same type signals based on measurements. In order to obtatin an accurate model, two signal selection approaches are proposed:a linear correlation coefficient index and a correlation coefficient map is proposed to describe the correlation between different measurement locations with geographic information; on the other hand, a nonlinear geometric template matching (GeTeM) and quality-threshold (QT) clustering are employed to compute the similarity of frequency dynamic responses in different locations of the power system. The complexity can be reduced and the accuracy can be improved by using these two proposed approaches. All the proposed approaches are verified by the simulation and actual data.(3) An adaptive wide-area damping control system (WADCS) based on solely wide-area measurements are proposed and implemented on a hardware test-bed (HTB) platform. A multi-input multi-output (MIMO) system model was identified to represent the main oscillation behaviors and a wide-area damping control loop selection approach was proposed based on this MIMO system model. To achieve fast online implementation, the identified high order MIMO model is deformed into a low order single-input single-output (SISO) model for the domainate oscillation mode to reduce the complexity of the model. Furthermore, an adaptive WADCS considering time delays was designed and the parameters can be updated online with the prediction dominate mode by compensating the time delay locally. In the end, online implementation in a two-area four-machine system on the HTB demonstrates the feasibility of practical realization of a measurement-based WADCS over a wide range of operation conditions. |
PDF Full Text Request