Open AccessIntelligent islanding detection method for photovoltaic power system based on Adaboost algorithm
Author(s) -
Ke Jia,
Zhengxuan Zhu,
Zhe Yang,
Yu Fang,
Tianshu Bi,
Jiankang Zhang
Publication year - 2020
Publication title -
iet generation, transmission and distribution
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.92
H-Index - 110
eISSN - 1751-8695
pISSN - 1751-8687
DOI - 10.1049/iet-gtd.2018.6841
Subject(s) - islanding , photovoltaic system , adaboost , boosting (machine learning) , computer science , algorithm , electric power system , artificial intelligence , power (physics) , pattern recognition (psychology) , electronic engineering , support vector machine , engineering , quantum mechanics , electrical engineering , physics
The universal islanding detection methods (IDMs) for photovoltaic (PV) power systems require manually thresholds setting. That will lead to a certain non‐detection zone (NDZ). Moreover, disturbance signals injected by active detection methods may adversely affect power quality. Aiming at the above problems, this study proposes a passive intelligent IDM for parallel multi‐PV system based on improved Adaptive Boosting (Adaboost) algorithm. Using Adaboost algorithm to generate classification models for islanding detection can theoretically avoid the NDZ of passive methods. The proposed method takes advantage of the electrical connection between characteristic parameters to adjust the classification model and improves the detection ability by redistributing the weight of each sub‐model. Simulation results show that when adopted to a multi‐PV system, the proposed method can effectively distinguish islanding operation in the NDZs of conventional passive IDMs. The method can also achieve accurate detection in the case of short‐term power quality interferences, line faults and disturbance signal interference injected by active methods.