电子测量与仪器学报

张朝龙,何怡刚,杜博伦,张兰芳,江善和.基于深度学习的电力变压器智能故障诊断方法[J].电子测量与仪器学报,2020,34(1):81-89

基于深度学习的电力变压器智能故障诊断方法

Intelligent fault diagnosis method of power transformer using deep learning

DOI：

英文关键词:transformer fault diagnosis self powered radio frequency identification SAE WNB

基金项目:国家自然科学基金项目（51607004,51577046，51777050）、国家自然科学基金重点项目（51637004）、国家重点研发计划“重大科学仪器设备开发”（2016YFF0102200）、装备预先研究重点项目（41402040301）、安徽高校自然科学研究重点项目（KJ2018A0369）资助

作者	单位
张朝龙	1.安庆师范大学物理与电气工程学院,2.武汉大学电气与自动化学院
何怡刚	2.武汉大学电气与自动化学院
杜博伦	2.武汉大学电气与自动化学院
张兰芳	1.安庆师范大学物理与电气工程学院
江善和	1.安庆师范大学物理与电气工程学院

Author	Institution
Zhang Chaolong	1. School of Physics and Electronic Engineering, Anqing Normal University, 2. School of Electrical Engineering and Automation, Wuhan University
He Yigang	2. School of Electrical Engineering and Automation, Wuhan University
Du Bolun	2. School of Electrical Engineering and Automation, Wuhan University
Zhang Lanfang	1. School of Physics and Electronic Engineering, Anqing Normal University
Jiang Shanhe	1. School of Physics and Electronic Engineering, Anqing Normal University

摘要点击次数: 1028

全文下载次数: 0

中文摘要:

针对电力变压器的故障诊断问题，提出了一种可用于海量监测数据的智能故障诊断方法。首先设计了无源射频识别（radio frequency identification, RFID）传感器标签用于测量变压器的振动信号，该设计具有结构简单、便利性强和功耗低等优点。针对于测量的变压器振动信号数量大、维度高、成分复杂、信噪比低等特点，利用深度学习技术中堆叠自编码器对测量的信号进行特征提取，提取的特征具有相同状态高度聚集，不同状态明显分离的优点。随后，基于提取的海量特征数据，应用加权贝叶斯分类模型进行故障诊断。为进一步提高故障诊断方法的性能，提出了混沌量子粒子群算法分别对堆叠自编码器和加权贝叶斯分类模型的参数进行优化。通过一个10 kV变压器的故障诊断实验表明，设计的无源RFID传感器标签能可靠地获取变压器振动信号，提出的故障诊断方法具有较高的故障诊断正确率。

英文摘要:

An intelligent fault diagnosis method used for magnanimous monitor data is proposed to deal with the problems of power transformer fault diagnosis. Firstly, a self powered RFID sensor tag is designed to measure the transformer vibration signals, which has advantages of simple structure, convenience and low cost. The measured transformer vibration signals have characters of large quantity, high dimension, complex components and low signal to noise ratio. A stacked autoencoder (SAE) of deep learning is employed to extract features of the measured vibration signals, where features of the same status are highly aggregated and features of the different statuses are obviously separated. A weighted native bayes (WNB) classification model is employed to the transformer fault diagnosis based on the extracted magnanimous feature data. To further improve the performance of fault diagnosis method, chaotic quantum behaved particle swarm optimization is proposed to optimize the parameters of SAE and WNB classification model, respectively. A 10 kV transformer fault diagnosis results show that the proposed RFID sensor tag can reliably collect the vibration signals, and the fault diagnosis method has a high correct rate of fault diagnosis.

查看全文查看/发表评论下载PDF阅读器