Within a wide-area electricity system inter-area oscillations occur as oscillations between the rotors of generator groups. System immanent power imbalances excite these oscillations, which result in oscillations of the grid frequency as well as in oscillating power flows. These power flows can lead to an overload of individual lines, separation of the grid, or even the loss of subnetworks.
Real-time monitoring of inter-area oscillations is rarely applied in the current system of the European Transmission System Operators. Neither are potential oscillations considered in the preventive grid security analyses. Besides other reasons this is due to the fact that there is no method available which can guarantee a robust and precise monitoring. This thesis presents a method which enables the identification of the crucial parameter of an inter-area oscillation, the periodic time and the damping ratio.
The thesis compares existing estimation methods. The combination of the two most promising methods, the Kalman filter process and the wavelet transformation method, leads to a method enabling real time monitoring and a robust estimation of the damping ratio. Known methods are all lacking this ability especially in ambient conditions. The functioning of the new method is applied on real measurement data of an undamped inter-area-oscillation of the year 2011.
Based on a grid model of the Continental European power system the dominant oscillations modes are identified. With the knowledge of the main oscillation modes a monitoring system with minimum number of measurement units is determined.
Available frequency measurement data from the monitoring system of the University of Stuttgart facilitate the estimation of periodic time and damping ratio for the main modes of the year 2013. Based on public data of the electricity grid, grid load, infeed of renewables and the power exchange between the countries, a regression analysis is conducted. The analysis assesses the impact of the given data on the substantial parameters of inter area oscillations. A strong correlation between the residual grid load and the periodic time of the oscillations modes can be identified. The correlation of the damping ratio with the residual load is, depending on the respective mode, either positive or negative and less strong as the correlation with the periodic time. A multiple regression analysis based on all related public data increases the precision of the prediction. The periodic time can be forecasted with an accuracy of 68 % and the damping ratio with an accuracy of 31 %.