摘要
以三电平光伏逆变器作为研究对象,基于小波包能量谱基数对其故障情况提出一种快速诊断方式。在小波包技术下对三电平逆变器桥臂电压及相应信号产生的能量谱特征向量进行分解分析,随后在极端学习机的诊断下,对其内部单或多个器件故障模式进行分析并排除。研究表明,提出的小波包能量谱技术具有较强的技术优越性,相比于传统模式下的BP神经网络以及最小二乘法故障诊断方式,该方式对于检测信号的获取更为容易,拥有较高的抗干扰性,其诊断速度较快,对于故障能够以精确的方式得到原因,从而显著降低故障诊断成本,是较为适宜的在线诊断方式。
In this paper,three-level Solar inverter was selected as the research object,and a fast diagnosis method was proposed for its fault based on the wavelet packet energy spectrum base.Under the wavelet packet technology,the energy spectrum eigenvector generated by the three-level inverter leg voltage and corresponding signals was decomposed and analyzed,and then under the diagnosis of extreme learning machine,the fault mode of its internal single or multiple devices was analyzed and eliminated.After relevant research and analysis,it can be concluded that the wavelet packet energy spectrum technology proposed in this article has strong technical advantages.Compared with traditional BP neural networks and least squares fault diagnosis methods,this method is easier to obtain detection signals,has greater antiinterference ability and a fast diagnostic speed which can accurately identify the cause of faults,thereby significantly reducing the cost of fault diagnosis,suitable for online diagnosis.
作者
曾瑞江
黄缙华
李志勇
ZENG Ruijiang;HUANG Jinhua;LI Zhiyong(Key Laboratory of Power Automation of China Southern Power Grid Corporation,Electric Power Science Research Institute of Guangdong Power Grid Co.,Ltd.,Guangzhou 510630,China)
出处
《电工材料》
CAS
2024年第4期71-74,共4页
Electrical Engineering Materials
基金
南方电网公司科技项目资助(GDKJXM20210066)。
关键词
小波包能量谱
光伏逆变器
故障诊断
ELM
wavelet packet energy spectrum
Solar inverter
fault diagnosis
ELM
