Transient Stability Analysis of Large Scale Power Systems Using Energy Methods.

Abstract

The transient stability analysis of large scale power systems is an important problem in power system planning and operation. Social and economical factors have forced electrical utilities to use electric energy from remote baseload power plants far removed from the customers. This has resulted in an increasingly complex system of grid interconnections that must operate closer to the margin of stability. Currently the commonly used method for determining the system stability is the direct simulation of the system. One alternative to direct simulation that has attracted many researchers is Lyapunov's Direct Method or the energy method. While some of these reseachers have "fine tuned" this method to remove its conservative nature, there still exist some major limitations to the approach. The objective of this research is to answer some fundamental questions in this area to make this approach a viable alternative to direct simulation. Three important achievements of the research conducted are: (1) A new energy function has been derived that accounts for the system voltage fluctuations. This results in a more accurate determination of the boundary of stability. (2) A highly efficient algorithm based on maxflow algorithm has been developed that can be used for the determination of stability limits. This algorithm does not require one to determine all the equilibria of the system and then select the critical one. The efficiency of the proposed approach arises from the fact that one can directly identify and compute the critical equilibrium value. (3) Closely related to the issue of stability is the first swing or short term stability. In this, one is primarily concerned with the system trajectory for a short period of time following the fault. A general approach has been developed for short term stability analysis. In this approach, which is a special case of Lagrange Stability, approximate "trajectories" of the system can be easily evaluated; with these approximate "trajectories" it is possible to estimate the time after which the system may become unstable. Unlike the energy method, this method places no restriction on the line losses or load models.