Transient stability of power system-a survey

This brief survey attempts to overview several available methods for assessing the transient stability of a power system. There are six methods considered and discussed in this study. They can be classified into three bases, i.e. digital simulation (numerical integration method, direct or Lyapunov methods, and probabilistic), heuristic (expert system), and training (pattem recognition and artificial neural network). The merits and potentialities and the drawbacks of these approaches are described and then compared. The frst based approaches are classified as the conventional transient stability analysis. These approaches usually use very extensive mathematical models of the system. Digital computer simulation with large computational time is needed in solving the problem with these approaches. The heuristic approach is an artificial intelligence (AI) approach. Human expertize is encoded in the rule-based program. This, along with a set of numerical algorithms, constitute an expert system. The training based approaches use off-line generation of pattern sets to train the assessors. The difference between these two is the fact that the artificial neural network (ANN) is able to learn and then generalize the pattem after training stage. ANN has also been used in the last five years in the areas of power systems. It shows its promising potentialities for on-line monitoring. The present achievements on this approach and its future trends are explored and discussed.