An online contingency filtering scheme for dynamic security assessment

This paper details the philosophy and the development of a cascade of progressively restrictive contingency filters to classify and rank contingencies for dynamic security assessment. Dynamic security assessment deals with the determination of contingencies which cause power system limit violations (e.g., unacceptable transient voltage dips) and/or system instability. It is assumed that (Figure Presented) Figure 1. Conceptual outline of the filtering scheme dynamic security of a power system subjected to large disturbances is assessed. The need for performing dynamic security analysis in near real time is highly recognized. For this to be made possible, a smaller number of possible critical contingencies must be identified and analyzed to determine the safe regimes of operation. This is the motivation for this work. Also, this methodology is not being used to address the issue of stability assessment. The purpose is to identify all possibly critical contingencies for more detailed analysis. The approach presented here uses the STEF method to select contingencies which are likely to cause dynamic security violations. Severity ranking and analysis of contingencies is thought of as a filtering process where the most severe cases are identified and ranked high on the list for more exact analysis. The nonsevere cases are filtered out of the list. A cascade of progressively restrictive contingency filters is used. The conceptual outline of the filtering scheme is shown in Figure 1. All the filters use detailed machine models with excitation control. Also, the modeling includes HVDC and nonlinear loads. The various levels of filters use different approximations in the calculations of the STEF method. The filtering scheme is tested on a Northern States Power (NSP) test system. The results of the filtering scheme are compared against those obtained using a conventional time domain simulation program (EPRI-ETMSP) to demonstrate the efficacy of the filtering concept.