摘要
提出一种基于小波时频图和改进型LeNet-5卷积神经网络的汽轮机转子系统故障诊断方法。首先利用柔性转子实验台模拟汽轮机转子系统各种运行状态,采集转子系统在各种运行状态下的振动信号,然后采用连续小波变换将振动信号转换为二维时频图像,最后利用卷积神经网络优异的图像识别能力实现转子系统的故障诊断。实验结果表明:提出的方法能够有效地实现转子系统故障诊断,具有较高的故障识别准确率。
In this paper,we propose a fault diagnosis method for turbine rotor system based on wavelet time-frequency diagram and improved LeNet-5 convolutional neural network.Firstly,the vibration signals of the rotor system are collected under various operating conditions by simulating the turbine rotor system with a flexible rotor test bench,and then the vibration signals are converted into two-dimensional time-frequency images by using continuous wavelet transform.The experimental results show that the method proposed in this paper can effectively achieve rotor system fault diagnosis with high fault identification accuracy.
作者
高玉才
付忠广
谢玉存
王诗云
杨云溪
GAO Yu-cai;FU Zhong-guang;XIE Yu-cun;WANG Shi-yun;YANG Yun-xi(Key Laboratory of Power Station Energy Transfer Conversion and System,North China Electric Power University,Ministry of Education,Beijing 102206,China)
出处
《汽轮机技术》
北大核心
2021年第6期445-447,454,共4页
Turbine Technology
基金
国家自然科学基金(50776029)。
关键词
转子系统
振动信号
小波变换
故障诊断
rotor system
vibration signal
wavelet transform
fault diagnosis
