Reduction of Dynamic models of Power System

Abstract: J.E. Van Ness, H. Zimmer and M. Cultu PAS 93. No.1, January/February 1974 p.3

The order of the differential equations required to represent in detail the dynamic response of a large power system with its associated voltage regulators and governor is so high that some degree of approximation is almost always necessary. This is usually done by neglecting those part of the block diagram with small time constant or in some cases by neglecting the voltage regulators of governors completely. More exact method that have been proposed retain the dominant eigenvalues and corresponding eigenvectors in the reduce model. This gives and more accurate representative, but interconnections are introduced trough out the model which do not exist in the physical system. Consequently, the resulting system matrix usually has very few, if any, Zeros.

In the paper the values of some element of the eigenvectors are relaxed in order that the topology of the resulting models can be specified. This is the same as specifying that certain entries of the system matrix be zero. Some conditions for the existence of this new model are given. Several numerical methods for finding the reduced model with the topology specified are presented, and their relative merits are discussed. Examples of the application of this method to actual systems are also presented.