| The objective of this dissertation is to present methods to characterize and predict rotor angle stability in wide area power systems and to find ways to control and enhance transient stability. Three approaches are presented in this dissertation for characterizing and predicting rotor angle stability.; The first approach for characterizing rotor angle stability in wide area power systems is based on a new screening and tracking algorithm. This algorithm tracks the time domain relative rotor angles using six criteria. These criteria are the first swing, second swing with increasing rotor angle, increasing oscillatory rotor angle, first swing with decreasing rotor angle, second swing with decreasing rotor angle, and oscillatory decreasing rotor angle.; Based on the screening and tracking algorithm, generator tripping look-up tables are proposed. The method uses off-line time domain analysis to build look-up tables of required generator trip aiming. The method employs two agents. The first agent determines if generator tripping is required to attain transient stability for each credible system disturbance and operating condition. Tracking of relative rotor angles is used in the first agent. The second agent compares the simulated electric power for each generator before and after the disturbance to determine the most appropriate generator tripping for transient stability.; In the second approach, a multi-agent technique is presented for the preparation of look-up table for arming of fast valving. This technique is divided into two parts. The first part is an off-line prediction agent that will predict power system instability for simulated scenarios. According to the prediction agent's output, a control agent is identified. The control agent is turbine fast valving. Fast valving of one or more units within the area where the fault occurs minimizes the accelerating power and hence reduces the transient power swings. As a result, the full or partial power output is restored within a few seconds.; The third approach for characterizing and predicting rotor angle stability is based on the use of the associate Hermite functions. The relative rotor angle data are used for building a model by an associate Hermite series expansion. The coefficients of the model are estimated using least mean squares (LMS). The built model is used for predicting the relative rotor angles. Continuous updating of the model parameters is carried out to ensure that the accuracy of the predicted data is within a predetermined limit. |
