WT-ANFIS在孤岛检测中的应用研究

针对传统被动式孤岛检测法存在检测时间长、盲区大,而主动式孤岛检测法影响电能质量的缺点,提出一种新的基于模糊神经网络与小波变换的孤岛检测方法。该方法首先采集逆变器输出的电流信号和公共耦合点处的电压信号,再将该电流信号和电压信号分别进行小波变换,然后通过对各尺度上的细节信号进行算法处理来获取适合于孤岛检测的特征向量,最后该特征向量通过模糊神经网络进行模式识别来判断系统是否发生孤岛现象。仿真与实验结果表明,该方法在并网逆变器功率与本地负载功率匹配及失配的多种条件下均能有效识别,具有检测速度快,盲区小,对电能质量无影响等优点,并且适合于单相、三相光伏并网系统。

Applied research of WT-ANFIS in islanding detection

The detecting time is long and non-detection zone ( NDZ) is large for traditional passive islan-ding detection methods, while active methods have some negative effects on power quality. A novel islan-ding detection method was proposed based on wavelet transform ( WT) and adaptive network based fuzzy inference system ( ANFIS) . Firstly, output current of inverter and the voltage of point of common cou-pling ( PCC) were gathered, and then WT was adopted to analyze the current signal and the voltage sig-nal. Secondly, the detail signals on all levels were used to extract characteristic vectors. Lastly, ANFIS used these characteristic vectors to pattern recognition and determine whether there was an island phe-nomenon. The simulation and experiment results show that the proposed method has the advantages that detecting time is short and non-detection zone ( NDZ) is small and effectively identify all kinds of load conditions such as the power of grid-connected inverter match and mismatch the one of local loads,and can be used for single-phase and three-phases photovoltaic grid-connected system.