A tracking time domain simulator for real-time transient stability analysis

Real-time power system transient stability analysis requires the analysis of hundreds of contingencies in terms of minutes using online data from state estimation. The final objective is to present timely information about transfer limits and stability margins and eventually implement corrective actions. In this paper, the authors assume that the dynamic contingency analysis (DCA) has to be repeated every 15 minutes. During this period of time, the loading and configuration conditions of the system change significantly but not drastically. They verify that the set of the relevant contingencies remains almost the same in the time interval comprised between two cycles of the functions associated to dynamic security analysis (DSA). Parallel-in-time formulated algorithms can be used in tracking the load conditions in real-time adopting as initial guess for the simulation of each contingency, the trajectories obtained for the same contingency at the previous cycle of DSA functions (that is the ones calculated in the previous 15 minutes). In this way, the information obtained in a previous cycle of DCA is not lost completely as in step-by-step algorithms. The feasibility of the approach has been tested on the Italian 380-220kV transmission system operated by ENEL. A parallel implementation of the approach on a nCUBE multiprocessor is reported.