基于BP神经网络的光伏系统故障诊断方法

A Fault Diagnosis Technique for Photovoltaic System Based on BP Neural Networks

在恶劣多变的环境下,光伏发电系统易发生多种故障,简单的监控及故障诊断技术无法实现系统的智能化和信息化。为此,提出一种基于Levenberg-Marquardt算法的BP神经网络(BPNN)故障诊断方法。获取光伏阵列的输出电压与电流、逆变器输出电压、直流负载电压、辐照度和温度数据,通过LM-BPNN算法挖掘出运行数据与故障模式之间的隐含映射关系,从而识别出光伏发电系统多种故障类型。最后,通过自制光伏电站模拟平台验证了方法的有效性。

Photovoltaic power generation system takes place to malfunctions under the influence of harsh and volatile environment,so simple fault diagnosis and monitoring technology can＇t realize the intelligent and informatization of power generation system.For this reason,an improved fault diagnosis method based on Levenberg-Marquardt（L-M）algorithm optimized BP neural networks（BPNN）is proposed in this paper.The proposed method selects the total output voltage of photovoltaic array,the total output current of photovoltaic array,the inverter voltage,the load voltage,the overall irradiance and the environmental temperature to mine various operation data and failure modes of the implicit mapping relationship through the LM-BPNN algorithm,and identify the fault types of photovoltaic power generation system.Self-made photovoltaic power plant simulation platform simulates fault conditions and the validity of the proposed method is verified experimentally.