High Information Content Database Generation for Data Mining based Power System Operational Planning Studies

被引：0

作者：

Krishnan, Venkat ^{[1
]}

McCalley, James D. ^{[1
]}

Henry, Sebastien ^{[2
]}

Issad, Samir ^{[2
]}

机构：

[1] Iowa State Univ, Dept Elect & Comp Engn, Ames, IA 50011 USA

[2] RTE, Versailles, France

来源：

IEEE POWER AND ENERGY SOCIETY GENERAL MEETING 2010 | 2010年

关键词：

Monte Carlo Simulation; Information Content; Decision Tree; Importance Sampling; Voltage Stability;

D O I：

暂无

中图分类号：

TE [石油、天然气工业]; TK [能源与动力工程];

学科分类号：

0807 ; 0820 ;

摘要：

Database generation for training is a critical aspect of the performance of data mining based power system reliability studies. Traditionally, Monte Carlo sampling of operational parameters are done to form various basecases and contingency analysis is performed to obtain the training database. This paper proposes an efficient sampling strategy that maximizes information content while minimizing computing requirements to form a training database for decision tree based operational planning studies. A Monte-Carlo variance-reduction method, namely importance sampling, is used to construct the proposed sampling approach. The method developed is tested on the Brittany area of RTE's system for a voltage stability assessment study, and decision rules are shown to have improved accuracy.

引用

页数：8