Skip to Main Content
This paper describes new matrix transformations suited to the efficient calculation of critical eigenvalues of large scale power system dynamic models. The key advantage of these methods is their ability to converge to the critical eigenvalues (unstable or low damped) of the system almost independently of the given initial estimate. Matrix transforms such as inverse iteration and S-matrix can be thought of as special cases of the described method. These transforms can also be used to inhibit convergence to a known eigenvalue, yielding better overall efficiency when finding several eigenvalues.