A NEW APPROACH TO DYNAMIC CONTINGENCY SELECTION

Contingency selection is important for the security analysis of a large scale power system since it can reduce the required computational effort by ranking the various outage events according to their severities. Among the three aspects of power system security, i.e., transient security, dynamic security, and steady-state security, contingency selection for dynamic security assessment is of concern in this paper. The proposed method is to use an iterative method to compute the eigenvalues of the system under outage conditions. The initial values for the iterative procedure are the eigenvalues for the normal operating condition (base case eigenvalues). To reduce the computational burden, the eigenvalues of the system under outage conditions from the first iteration are employed for contingency ranking. From the results obtained from the study on Taiwan power system, it is concluded that contingency ranking using the eigenvalues from the first iteration is both accurate and efficient. It is also found that only the eigenvalues for the worst-damped mode must be considered in contingency ranking. To further improve the efficiency of the proposed method, results obtained from the first iteration of the fast decoupled load flow (FDLF) are also used to model the operating conditions after contingencies. Again, it is observed that satisfactory dynamic contingency selection can be achieved by using the first iteration of FDLF with much less computational effort than that required by full AC power flow. It is thus concluded that accurate dynamic contingency selection can be performed efficiently by first computing the operating conditions after contingencies using the first iteration of FDLF and then calculating the eigenvalues for the worst-damped mode by using the first iteration of eigenvalue computations