The accuracy with which power system transient stability limits can be calculated by computer simulation depends on the synchronous machine model. The saturation representation in the machine model can be improved now by non-linear magnetic finite element analyses of the machine. This paper describes steps taken to incorporate a saturation algorithm, based upon finite element analyses, into an 1100 bus power system stability program. Power transfer capability of a hypothetical system is compared using the finite element based algorithm versus the conventional algorithm based upon the manufacturer's open circuit saturation curve.
INTRODUCT [ON: BACKGROUND AND MOTIVATIONSimulation studies are made to calculate stability limits of power systems as part of the quantitative analyses of alternatives facing system planners and operators.The alternatives include the generation dispatch and consequent amount of power transfer among and within utilities.Synchronous generators are modeled within these studies, to represent the nonlinear dynamics of the conversion from rotating shaft power to three-phase electric power.Prior work funded by EPRI in finite element analysis of generator modeling and validation by test (1,2,3] has shown that finite element analyses may be used to prepare generator models from design data, and that those models are well suited for calculation of the dynamic response (1,3] and steady-state conditions of a generator [1,2]. These results hold promise of improving the accuracy of power system analyses made by computer simulation.Finite-element based (FE) dynamic generator models (i.e., using linear operational inductances instead of the traditional T' , T'O, etc.) were developed in RP1288 [1,3]. The F dyn%mic models were used in simulation analyses of staged stability tests at the Bridger and Lambton stations; the comparisons indicate that the FE dynamic models are as accurate as models derived from on-line test methods [4]. That is, both FE models and models derived from tests appear to be more accurate than those based upon manufacturer's data (which are in turn based upon IEEE Standard 115 (5]). The major advantage in using FE dynamic models *A companion paper,
