多逆变器并网条件下新型孤岛检测方法研究

多逆变器并网孤岛检测方法包括被动式孤岛检测和主动式孤岛检测两大类。文章首先分析了一种主动式多逆变器并网孤岛检测方法，即注入负序扰动电流的孤岛检测法。针对该方法的缺点，提出了一种新型被动式多逆变器并网孤岛检测方法，即小波技术与神经网络相结合的孤岛检测法。该方法先利用小波变换技术获取相应的信号特征，再利用神经网络根据其获取的信号特征对分布式发电系统的孤岛或非孤岛工作状态进行辨识。仿真结果表明，该方法有很好的孤岛检测性能，且检测时间短，同时不会影响系统的稳定性和供电质量。

Research on a New Islanding Detection Method Under the Condition of Multiple Grid-Connected Inverters

Islanding detection methods for multiple grid-connected inverters include passive and active islanding detection. This paper first analyzes an active islanding detection method, which is conducted by injecting negative sequence disturbance current. Considering the shortcomings of this method,a new passive multiple grid-connected inverter islanding detection method is proposed,which combines wavelet technology and neural network. In this method,the wavelet transform technology is used to obtain the corresponding signal features, and then the neural network is used to identify the islanding or non-islanding working state of the distributed generation system according to the obtained signal features. The simulation results show that this method has excellent islanding detection performance,short detection time,and does not affect the stability and power quality of the system.