State-space system identification-toward MIMO models for modalanalysis and optimization of bulk power systems

摘要

This paper provides an introduction to a reduced-order,nsmall-signal identification approach to modal analysis and control ofnlarge power systems. Being based on system-wide responses to low-energynpulse excitations generated using conventional time-domain simulationnsoftware such as PSS/E or EMTSP, it readily takes full advantage of thenlarge built-in model database. The proposed multi-input-multi-outputn(MIMO) minimal realization reveals naturally the dominant modes attachednspecifically to a given device, as well as the transfer functionsnrelating selected measurement and observation sites. It plays ancomplementary role to direct computation of the full-scale linearizednmodel using a comprehensive program such as MASS, after a summary of thentheoretical work initiated at Hydro-Quebec in the early 1990s to promotenthis approach and put it into routine use, we present the mainnchallenges in developing a production grade computer code. Detailednexamples inspired by actual network studies at Hydro-Quebec arendiscussed, the most complex of them involving the identification of an125th order MIMO model with 26 inputs and 26 outputs