An artificial intelligence system for power system contingency screening

The power system security assessment problem is considered. The objective is to investigate the extent to which a computer program with knowledge and inference capabilities can assist a human operator in this task. A rule-based system for screening the consequences of single outages was developed which can be extended to deal with multiple outages. The rules are independent of the size of the power system and derive partly from human operator expertise and partly from simulation models. Results show that such systems can be very effective in focusing attention on a much-reduced number of the more critical contingencies. The expert system was demonstrated using a power system model of moderate size (25 buses, 35 lines). This approach is cost effective and can be applied to online operation.>