Study Of Algorithm For Power System Dynamic Voltage Stability Control

The large-scale wide area interconnected electric network and deregulated market environment of power systems have driven the operation of power systems closer to their limits than ever, and urge the electric power industry to study and develop more effective stability control methodologies. This thesis involves two classical research fields, namely, voltage and angle stability, and focuses on the power system stability control. The main contents include the voltage stability control based on the operation state of power systems and angle stability control based on the FACTS controller. The following are the main contributions of this thesis.This thesis proposes a novel model for optimal coordinated voltage emergency control. It takes account of the dynamics of loads and discrete/continuous nature of controls with coordination of dissimilar controls at different geographical locations in order to prevent voltage collapse during an emergency. An integration index of the load bus voltage deviation is adopted as the voltage stability performance index.Based on the optimal control theory, the sensitivities (gradients) of this performance index with respect to controls are derived rigorously. The sensitivities can be evaluated by solving the adjoint equations using the fast quasi-steady-state (QSS) time domain simulation results. This novel model has the following advantages: Firstly, the availability of the sensitivity allows the complex intractable voltage optimization problem to be transformed into a classical quadratic programming problem. Secondly, load shedding is guided via the sensitivity information. The ones with higher sensitivity would have the higher priority to shed such that the minimum amount of load shedding can be ensured. Thirdly, this novel model is based on the QSS simulation; therefore it has the same adaptability to the system modeling and scale as the time domain simulation. Finally, the overall computational speed of the novel model is very competitive. Typical computation time of sensitivity is less than two times QSS computation time, and the computation speed of quadratic programming is much faster than the intelligence methods. Moreover, a method to select the buses vulnerable to voltage instability using the modal analysis technique along QSS unstable trajectory is also provided in this thesis.This thesis proposes a novel optimal preventive control methodology with voltage stability constraints. Compared to the existing preventive control methodologies, the proposed novel methodology do not require to identify the critical point along the QSS time domain trajectory of the post-contingency system. It can deal with not only the case that the dynamic trajectory of the post-contingency system is stable but voltages have violation, but also the case that the dynamic trajectory of the post-contingency system is unstable. Since this novel methodology is not dependent on the control type, it is convenient to extend the controls to other parameter spaces.The unified power flow controller (UPFC) is the most versatile FACTS controller, in this thesis the relative gain array (RGA) method in control theory is first applied to analyze the interaction degree among UPFC multiple control channels, and select the best variable pairs for UPFC under which the interaction among UPFC control channels is the weakest. Furthermore, the influences of operation mode variations and oscillation frequency variations on best pairs are analyzed. Based on RGA method this thesis for the first time provides the rigorous theory support for the widely used control scheme of UPFC. Novel control strategies of DC capacitor voltage are then proposed and lead to the design of a novel two-stage control scheme for UPFC to damp tie-line low frequency oscillation. Moreover, an effective estimation method based on the locally measurable quantities is proposed in this thesis to estimate the angular speed difference between centers of inertia (COI), and the estimation result is quite accurate.