In order to ensure good power quality of photovoltaic power generation system and safety of equipment and physical .For traditional islanding detection method NDZ (Non-detection zone) is big and the introduction of a large disturbance reduce power quality, this paper presents a wavelet entropy and achieve a new algorithm uses adaboost islanding detection method by sampling the voltage of the common point and the inverter current extracted from the corresponding wavelet entropy feature quantity after the wavelet transform, and then use Adaboost-svm improved training and recognition algorithms, simulation results show that the use of this article the method can accurately identify very difficult to identify the traditional methods of island situations, and in the same learning algorithm, the accuracy rate is relatively high, with high accuracy and reliability in islanding detection.

 Island detection, multi resolution, wavelet entropy, Adaboost-SVM

[1] Xie Dong, Zhang Xing, Cao Renxian.Automatic island detection based on wavelet transform and neural network [J]. Proceeding of the CSEE, 2014 (4): 537-544.

[2] ZHU Yan-wei, SHI Xin-chun, LI Peng.Application of multi-resolution singular spectral entropy and support vector machine in islanding and disturbance identification [J]. Proceeding of the CSEE, 2011 (7): 64-70. 4

[3] Wang Xiaodan, Sun Dongyan, Zheng Chunying, et al.A SVM classifier based on AdaBoost [J] .Journal of Air Force Engineering University: Natural Science Edition, 2006, 7 (6): 54-57.

[4] ZHANG Pei-chao, TAN Xiao-feng, YANG Pei-xin.Key feature recognition and meta-learning method for island detection [J] Power System Automation, 2014, 38 (18): 72-78.

[5] Wang Guosheng. Support vector machine theory and algorithm research [D]. Beijing University of Posts and Telecommunications, 2008.

[6] Xie Ping, Liu Bin, Lin Hongbin, et al. Multi-resolution singular spectral entropy and its application in vibration signal monitoring [J].

[7] HE Zheng-you, LIU Zhi-gang, QIAN Qing-quan.Discussion on the Wavelet Entropy Theory and Its Feasibility in Power System [J] .Power System Technology, 2004, 28 (21): 17-21.

[8] ZHANG Zhen, WANG Binqiang, LIANG Ningning, et al. A Traffic Classification Method Based on AdaBoost-SVM [J] .Application Research of Computers, 2013, 30 (5): 1481-1485.

[9] DONG Zhong-xing.An island detection model for microgrid based on feature extraction and classification algorithm [D]. North China Electric Power University, 2014.

[10] Redfern M A, Barrett J I, Usta O. A new microprocessor based islanding protection algorithm for dispersed storage and generation units[J]. IEEE Transactions on Power Delivery, 1995, 10(3):1249-1254.

[11] El-Arroudi K, Joos G, Kamwa I, et al. Intelligent-Based Approach to Islanding Detection in Distributed Generation[J]. IEEE Transactions on Power Delivery, 2007, 22(2):828-835.

[12] Wang Zhenyue. Photovoltaic grid-connected inverter islanding characteristics and detection methods [D]. North China Electric Power University, 2012.