Title :
Method for determining line drop compensator parameters of low voltage regulator using support vector machine
Author :
Kikusato, Hiroshi ; Takahashi, Naoyuki ; Yoshinaga, Jun ; Fujimoto, Yasutaka ; Hayashi, Yasuhiro ; Kusagawa, Shinichi ; Motegi, Noriyuki
Author_Institution :
JST, CREST, Waseda Univ., Tokyo, Japan
Abstract :
Highly accurate predictions of load demand and photovoltaic (PV) output have become possible in recent years because of improved measuring instruments and the increase of databases on load demand and PV output. The appropriate control parameters for actual power system operation can be determined by using these predictions. Although parameters determined by conventional methods are accurate, they may not be determined in time before the beginning of operation because extensive time is required for the calculations. In this paper, the support vector machine-a machine learning method that solves the two-class classification problem-is used to determine the line drop compensator (LDC) parameters instantly. To verify the validity of the proposed method, we carried out numerical simulations to determine the LDC parameters. From the simulated results, we found that the proposed method can instantly and accurately determine the LDC parameters.
Keywords :
learning (artificial intelligence); photovoltaic power systems; power distribution control; power distribution lines; power engineering computing; power generation control; support vector machines; voltage control; voltage regulators; actual power system operation; distribution systems; line drop compensator parameter determination method; load demand prediction; low voltage regulator; machine learning method; measuring instrument improvement; numerical simulations; photovoltaic output prediction; support vector machine; two-class classification problem; voltage control; Accuracy; Fluctuations; Photovoltaic systems; Support vector machines; Training data; Voltage control; Distribution systems; LDC method; LVR; SVM; Voltage control;
Conference_Titel :
Innovative Smart Grid Technologies Conference (ISGT), 2014 IEEE PES
Conference_Location :
Washington, DC
DOI :
10.1109/ISGT.2014.6816413