This thesis presents a novel approach to electromechanical oscillation damping estimation under the ambient conditions of a power system. The power system is said to operate under the ambient conditions when it is only subjected to ever present small excitations such as constantly varying load. The damping estimation method is based on the wavelet transform and the random decrement technique. The thesis reviews the properties of the wavelet transform that are essential in damping estimation, defines criteria for optimal mother wavelet selection in damping estimation, and identifies the possible mother wavelets in damping estimation of the Nordic power system. It also studies the optimal selection of other parameters and defines values for them. Both the simulated and measured power system data is analyzed with the damping estimation method in the thesis. The results show that when the parameter selections (especially the mother wavelet function and the time window length) of the damping estimation method are done correctly and a signal with good observability of the mode is used, the damping can be estimated close to the known damping of the simulation model using the ambient-excited oscillation data. The damping estimates are more accurate when the real damping of the oscillation mode is low; i.e. when it is more important to system stability. The measurement noise of the analyzed signals does not have much effect on the estimates. It is shown that the method can be used to observe the degraded damping due to incorrect operation of the power system damping controller. In addition, the method can possibly be applied to verification of the power system simulation model. Although the performance of the method is studied for the Nordic power system case, it is recognized that the method is applicable to other power system, too.
展开▼
