作者: A. Kaci , I. Kamwa , L.-A. Dessaint , S. Guillon
DOI: 10.1109/PESGM.2014.6939281
关键词: Random forest 、 Stability (probability) 、 Maximum power transfer theorem 、 Units of measurement 、 Phasor 、 Reliability (computer networking) 、 Engineering 、 Control theory 、 Power (physics) 、 Electronic engineering 、 Margin (machine learning)
摘要: In the United States, number of Phasor Measurement Units (PMU) will increase from 166 networked devices in 2010 to 1043 2014. According Department Energy, they are being installed order “evaluate and visualize reliability margin (which describes how close system is edge its stability boundary).” However, there still a lot debate academia industry around usefulness phase angles as unambiguous predictors dynamic stability. this paper, using 4-year actual data Hydro-Quebec EMS, it shown that enable satisfactory predictions power transfer security margins across critical interface random forest models, with both explanation level R-squares accuracy exceeding 99%. A generalized linear model (GLM) next implemented predict day-ahead hour-ahead time frames, historical values load forecast. Combining GLM based forecast mapping transfers result new data-driven approach for monitoring.