Inductive learning in power system voltage control

The application of inductive learning to voltage control and contingency assessment of power systems is considered. The first application discussed is the identification of correct control amount for remedial actions. The inductive learning application is intended to obtain a decision tree which identifies a proper value of the low voltage limit. It is shown that learning capability can be incorporated into a voltage control expert system (VCES) by including a decision tree. The VCES obtains the appropriate low voltage limit from a precomputed decision tree. The software implementation of th VCES with learning capability is described. The VCES with a learning module is integrated into the dispatcher training modulator environment. The second application is an attempt to use inductive learning to identify critical operating conditions/outages which may cause voltage problems in a power system. The selected attributes, generation of the training and test sets, and the numerical results are summarized