The Effects of External Modeling Errors on On-Line Security Analysis

The on-line security analysis function in a control center today processes the steady-state results of several hundred contingencies to periodically determine the security of the power system. For accurate results the analysis must be done on a representation of the total interconnected network. This requires the modeling of the external systems from which the control center does not receive real time data. In the absence of data links between control centers, the real time status and analog data for the external systems must be assumed or estimated. In this paper, the effects of errors in this assumed data on the security analysis results are presented. Using these effects as the criteria, the two most common external modeling methods are compared for their robustness by testing on two different power systems. The results show that status errors affect the security analysis results more than analog errors, that errors electrically closer to the internal system have larger effects than those further away, and that the reduced external models produce less accurate results than the more extensive models.