基于复变自适应神经网络的电网相位估计方法

针对非理想电网电压下,不平衡电压、频率偏移引起的电网相位难以检测的问题,提出了一种在复变域下使用的基于自适应神经网络的电网相位估计方法.首先,对非理想电网电压进行建模,在得到神经网络模型的基础上,将复变最小均方算法的权值更新方法应用到神经网络权值更新过程中,利用神经网络权值实现对相位的估计.为了跟踪电网频率,设计了电网频率跟踪环节,并对收敛性进行了分析.仿真和实验的结果表明所提出的方法能够快速准确地对非理想电压下的电网相位进行估计.

Phase Estimation Method for Power System Based on Complex Adaptive Neural Network

Aiming at the difficulty of grid phase detection under non-ideal voltage caused by unbalanced voltage and frequency fluctuation, a phase estimation method was proposed for power system based on complex adaptive neural network. On the basis of neural network model of non-ideal grid voltage, the weight update method of complex least mean squares to the process of weight update procedure of neural network was introduced. Then the weights of neural network were used to calculate the phase. To trace the frequency of grid, a frequency tracing unit was designed and it was proved to be convergent. The simulation and experiment results demonstrate that the proposed method is able to estimate the phase rapidly and precisely under non-ideal voltage conditions.