[C36] Semitekos D. D., Avouris N. M., Giannakopoulos G. B., A Flexible Machine Learning Environment for Steady State Security Assessment of Power Systems, Proc. 6th IASTED Int. Conference on Power and Energy Systems, pp. 190-194, July 2001 Rhodes, Greece, (ed.) P.B. Bourkas, ACTA Press, Anaheim, CA. (pdf)
Machine Learning techniques have been applied to power system analysis for a number of years. The need for a flexible computing environment to support these studies is derived by the complexity of the process, the volume of data often used and the diversity of the applied tools and techniques that span many disciplines. A data warehouse can be a central component of this environment. In this paper our experience with building and using such an environment is described. Data collection and transformation tools, machine-learning tools and testing tools have been integrated in the MLPSE environment described here, used for steady state security assessment of power systems. It is argued that the proposed approach can be applied to many similar power systems analysis studies.
Data warehouse, decision trees, machine-learning, steady state security assessment, contingency analysis.