Applications of compressed sensing and sparse representations for state estimation in power systems

Compressed sensing is an emerging signal processing method that finds applications in diverse estimation problems. This paper shows that power systems events may be analyzed in terms of sparse structures, especially if the probability of their occurrence within a power network is low. The paper presents several examples for such events, and suggests methods for locating them within large power systems using few measurements. Several types of sparse events are analyzed: faults, lightning strikes, polluting loads, and electricity thefts.