Reduced order dynamic equivalents for electric power systems : an analysis of the modal-coherent and slow-coherent equivalencing techniques

This thesis is the analysis and comparison of two methods for determining reduced-order coherent equivalents for use in power system transient stability studies; the modal-coherent and slow-coherent equivalencing techniques. This comparison is based upon the ability of each method to preserve the dominant modes of the original linearized model in the equivalent. Both equivalencing techniques use a two step process: group identification and group aggregation. The group aggregation method associated with the modal-coherent method is the inertial averaging method; the aggregation method associated with the slow-coherent method is the singular perturbations method.

The constituent matrix concept is used as a tool to compare each method on the same basis. Two example systems, the seven generator Michigan Electric Coordinated System (MECS) model and the ten generator New England System (NES) model, are used to demonstrate and compare the two methods. Since either aggregation procedure can be used with groups determined by the grouping algorithms, a two part analysis is used. The first part compares the grouping algorithms. The second compares the aggregation procedures.

An efficient method for the reduction of the ranking table determined by the modal-coherent method was developed.

The singular perturbations aggregation method associated with the slow-coherent technique was shown to be clearly superior to the inertial averaging aggregation method in preserving the dominant modes of the original system. The modal-coherent grouping algorithm was shown to be superior to the slow-coherent grouping algorithm in recognizing coherent groups in most cases.