Application of data mining techniques for automat settings in emergency control at Hydro-Qudbec

This paper describes a new approach used by Hydro-Quebec to optimize the settings of automatic devices installed in its main power plants to maintain secure operation under extreme contingencies. An example of application of this approach is given to illustrate how to apply data mining techniques for the setting of the automatic generator tripping and load shedding system (RPTC: rejet de production et teledelestage de charges in French) installed at the Churchill Falls hydroelectric power plant (5500 MW) in Labrador. Real time snapshots of the Hydro-Quebec power system collected over a 5 year period have been used to generate large amounts of results (database) by transient stability simulations. The database is processed by the data mining software developed by the University of Liege to construct the decision trees. This approach gives the most relevant parameters and finds optimal settings for the RPTC system at the Churchill Falls, minimizing the number of generator tripping while maintaining the same performance in terms of security coverage. New operation rules have thus been established.