基于自适应RBF神经网络的光伏阵列故障检测方法

A Fault Detection Method for Photovoltaic Array Based on Adaptive RBF Neural Network

摘要: 为应对光伏阵列故障检测的实际需求，设计了一种基于总体误差的自适应RBF神经网络算法。该算法依据网络当前训练过程的整体误差与上一次训练整体误差的对比结果，对学习速率和动量参数进行动态优化，让网络的学习方向更加清晰，有效提升了网络的学习效率。将基于总体误差的自适应RBF神经网络应用于光伏电站阵列的故障识别，借助Simulink软件平台搭建光伏阵列模型并获取不同故障状态下的运行数据，开展了针对性的故障检测实验。仿真结果表明，采用自适应RBF神经网络的故障检测方法能够实现对光伏阵列故障的高效、准确检测，展现出较好的应用价值。

Abstract: For the purpose of photovoltaic arrays fault detection, an adaptive RBF neural network algorithm based on overall error is designed in this paper. The learning rate and momentum factor are adjusted adaptively according to the overall error size relationship obtained from the last two training sessions of neural network. So, the learning course of neural network has clear directionality, and the learning efficiency of neural network is improved greatly. The adaptive RBF neural network based on overall error was used in photovoltaic array fault detection. A photovoltaic array model was built and fault data was obtained based on simulink firstly, and then fault detection experiments were conducted. The simulation results show that the fault detection method based on adaptive RBF neural network can achieve efficient and accurate detection of photovoltaic array faults. The detection method has good application value.