Oscillatory stability limit prediction using stochastic subspace identification

Determining stability limits and maximum loading margins in a power system is important and can be of significant help for system operators for preventing stability problems. In this paper, stochastic subspace identification is employed to extract the critical mode(s) from the measured ambient noise without requiring artificial disturbances (e.g., line outages, generator tripping, and adding/removing loads), so that the identified critical mode may be used as an online index to predict the closest oscillatory instability. The proposed index is not only independent of system models and truly represents the actual system, but it is also computationally efficient. The application of the proposed index to several realistic test systems is examined using a transient stability program and PSCAD/EMTDC, which has detailed models that can capture the full dynamic response of the system. The results show the feasibility of using the proposed identification technique and index for online detection of proximity to oscillatory stability problems.