Database-assisted frequency estimation for power system measurement

In modern power transmission systems, monitoring of transient processes becomes more important. Analyzing a transient process is hard since the process is not only transforming from one steady state into another, but it also includes event-driven oscillations from LC-circuits within the net. The number of such oscillations is unknown and not constant. Furthermore, the oscillations can have a small amplitude. The knowledge of such oscillations can help understanding what happens where in the net. To estimate these oscillations, we present a signal processing method based on subspace estimation and history-based data ranking. It enables the estimation of an a-priori unknown number of oscillations and so the rank, although their amplitude might be small compared to dominating components.