Fast Frequency-Domain Decomposition for Ambient Oscillation Monitoring

This paper proposes a multidimensional ambient oscillation monitoring algorithm denoted as fast frequency-domain decomposition (FFDD). Based on a new theoretical result, the algorithm is offered as an improvement over the previously proposed frequency-domain decomposition in that FFDD does not require time-consuming singular value decomposition and it does not require cross spectrum estimates. FFDD is useful for fast real-time ambient modal estimation of a large number of synchrophasor measurements. The algorithm is tested on archived event data from a real power system.