ON EQUIVALENT-CIRCUIT STRUCTURES FOR EMPIRICAL MODELING OF TURBINE-GENERATORS

This work is aimed at extending a discussion offered concerning a paper presented at the 1993 IEEE/PES Winter Meeting [1] which attempted to show that the ''differential leakage'' was not required and, in general, seriously questioned the need for a third-order model. Unfortunately, the simplifications in [1] were possible only using a per-unit system not compatible with industry standards and current engineering practices for saturation analysis. In this paper, mathematical reasoning is combined with maximum-likelihood estimation from actual SSFR data to show that, for network structures preserving the main-field magnetizing impedance, the differential leakage contributors significantly to dynamic response and proves mandatory for adequately modelling rotor impedance. Besides, the claim made in [1] that a third-order model may not be sufficiently parsimonious is not borne out by further verification of the author's own data and extensive experience with real SSFR data from a 1100-MVA turbine-generator.