| 基于小波包能量谱和ELM的光伏逆变器多故障在线诊断 |
| |
| 引用本文： | 姜媛媛,王友仁,吴祎,孙权,罗慧.基于小波包能量谱和ELM的光伏逆变器多故障在线诊断[J].仪器仪表学报,2015,36(9):2145-2152. |
| |
| 作者姓名： | 姜媛媛 王友仁 吴祎 孙权 罗慧 |
| |
| 作者单位： | 1.南京航空航天大学自动化学院南京210016；2.安徽理工大学电气与信息工程学院淮南232001；
3.南京农业大学工学院南京210095 |
| |
| 基金项目： | 国家自然科学基金（61371041）、江苏省普通高校研究生科研创新计划资助项目与中央高校基本科研业务费专项资金（CXLX11_0183）、国家自然科学基金（61401215）、航空科学基金（2013ZD52055）、国家商用飞机制造工程技术研究中心创新基金（SAMC14 JS 15 051）项目资助 |
| |
| 摘 要： | 以三电平光伏逆变器为研究对象,提出一种多故障模式快速诊断新方法。首先,利用小波包分解提取出三电平逆变器的桥臂电压和上、下管电压信号的能量谱特征向量,并利用主成分分析降维后获取故障特征向量;然后,基于极端学习机诊断模型分离出单器件及多器件开路等多种故障模式。实验结果表明,相比于传统BP神经网络、最小二乘支持向量机故障诊断方法,该方法检测信号易获取,抗干扰性强,诊断速度快、精度高,减小了诊断成本和复杂性,适用于在线诊断。
|
| 关 键 词： | 极端学习机 小波包分解 光伏逆变器 故障诊断 特征提取 |
Online multiple fault diagnosis for PV inverter based on wavelet packet
energy spectrum and extreme learning machine |
| |
| Jiang Yuanyuan,Wang Youren,Wu Yi,Sun Quan,Luo Hui.Online multiple fault diagnosis for PV inverter based on wavelet packet
energy spectrum and extreme learning machine[J].Chinese Journal of Scientific Instrument,2015,36(9):2145-2152. |
| |
| Authors: | Jiang Yuanyuan Wang Youren Wu Yi Sun Quan Luo Hui |
| |
| Affiliation: | 1. College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China;
2. College of Electrical and Information Engineering, Anhui University of Science and Technology, Huainan 232001, China;
3. College of Engineering, Nanjing Agricultural University, Nanjing 210095, China |
| |
| Abstract: | A novel rapid diagnosis method of multiple faults for three level PV inverters is proposed in this paper. Firstly, the wavelet packet decomposition is adopted to extract the voltage energy values of the bridge arm and upper, lower tubes in three level PV inverter as the energy spectrum feature vectors and the principal component analysis method is used to perform the dimension reduction of the wavelet packet energy feature vectors and obtain the fault feature vectors. Then, various fault modes, such as single and multiple device open circuit and etc. are isolated based on the extreme learning machine diagnostic model. Experiment results show that compared with traditional BP neural network and least squares support vector machine fault diagnosis methods, the proposed method has the advantages of easy to obtain the detection signal, strong resistance to interference, fast diagnosis speed and high precision, and can reduce the cost and complexity of fault diagnosis. The proposed method is also suitable for online diagnosis. |
| |
| Keywords: | extreme learning machine(ELM) wavelet packet decomposition PV inverter fault diagnosis feature extraction |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《仪器仪表学报》浏览原始摘要信息 |
|
点击此处可从《仪器仪表学报》下载全文 |
|
