Research On The Reduced-order Method For Small Signal Stability Of Multi-machine Power System Using Detail Model

With the building and operation of the project "electric power transmission from west to east" and the forming of the nationwide power system's inter-connection, the problem of power system small signal stability is more notable. The research standard on the small signal stability cannot adapt to the needs that are propitious to the enlarging scale of power system and the levels of stability analysis and control. To study the reduced-order analysis methods that not only can satisfy the need of the system scale but also have higher calculation precision has an important academic value and practical project sense. Based on the review and present tendency analysis for small signal stability research, in this thesis, the reduced-order analysis method for small signal stability of multi-machine power system using detail model is researched.In order to settle the contradiction among the dimension localization of QR, scale of power system and the precision of the models, in this thesis, a fuzzy block-decoupling method is brought forward to calculate the eigenvalues of multi-machine power system. This method realizes the reducing order calculation. Based on the electromechanical coupling characteristic of the multi-machine power system, the transitive closure is calculated depending on the synchronization torque coefficient matrix. Then a proper threshold value is chosen to achieve the reasonable block decoupling of multi-machine power system. Eigenvalues of block-decoupling subsystem are calculated respectively. The set of these eigenvalues can be regarded as the eigenvalues of primary system approximately. This method not only can control the scale of the block-decoupling subsystems, but also achieves the reduced-order eigenvalue calculation for multi-machine power system. So this method can avoid the dimension disaster in the QR and calculate all the eigenvalues of large power system with the effect of damping windings considered. Through analyzing the phenomenon of the pseudo complex value in the block-decoupling method, the pseudo complex value is considered as the result of eigenvalue variation because of the system being decoupled forcedly, but this phenomenon has no effect on the result of the small signal stability.The selective modal analysis method (SMA) based on the tradition equivalence second-order model achieves the reduced-order calculation for the electromechanical modes, but its calculation precision isn't higher and sometimes it may lose eigenvalues.Based on the exact equivalence second-order model of multi-machine using detailmodel of generator, an electromechanical calculation method is put forward. This method connects the fuzzy block-decoupling method and the selective modal analysis method. Firstly, all the electromechanical initial values with higher accuracy are provided by the fuzzy block-decoupling method, and then the exact eigenvalues are iterated by the selective modal analysis method. This method can not only avoid losing eigenvalues but also improve the calculational precision of electromechanical modes.The method of frequency domain analysis doesn't reach the applied requirement for project in the small signal stability of multi-machine power system. In order to settle this problem, in this thesis, the complete decoupling method for the small signal stability of multi-machine power system connects the complete decoupling model of multi-machine power system with the frequency domain analysis method to achieve the reduced-order analysis in frequency domain from three ways. The absolute and relative stability of multi-machine is analyzed intuitionally using the open-loop frequency characteristic charts and the performance index of open-loop frequency domain for each decoupling single-machine. The closed-loop main poles of each single-machine are attained approximately from the characteristic polynomial frequency characteristic of closed-loop system of each decoupling single-machine, and its exact value can be computed using iteration. In general, these values just are the...