摘要
针对非平稳信号时频分析,结合分数阶傅里叶变换和S变换,提出了一种分数阶S变换,将信号分析从时间-频率域推广到时间-分数阶频域,并应用于齿轮箱故障诊断。通过仿真信号和齿轮箱故障信号分析,验证了分数阶S变换良好的时频聚集性。采用脉冲耦合神经网络进一步提取分数阶S变换时频图的特征参数,对齿轮箱故障信号进行了分类。结果表明,基于分数阶S变换提取的特征参数能更有效地区分齿轮箱的不同状态信号,从而提高齿轮箱故障诊断精度。
For time-frequency analysis of non-stationary signals,combined with fractional Fourier transform and S transform,a fractional S transform was proposed.This fractional S transform extended traditional time-frequency domain to time-fractional frequency domain for signal analysis,and was applied to the gearbox fault diagnosis.Analysis results of the simulated signal and gearbox fault signals verified the good time-frequency aggregation of fractional S transform.Pulse coupled neural network was then employed for extracting feature parameters of fractional S transform time-frequency images.Gearbox fault signals from five gearbox states were classified by utilizing the extracted features.The results show that the extracted parameters based on fractional S transform can more effectively distinguish gearbox signals from different gearbox states,thereby enhancing the gearbox fault diagnosis accuracy.
作者
王红庆
Wang Hongqing(College of Science,Beijing Forestry University,Beijing 100083,China)
出处
《电子测量与仪器学报》
CSCD
北大核心
2019年第8期133-139,共7页
Journal of Electronic Measurement and Instrumentation
基金
国家自然科学基金(11501032)资助项目
关键词
时频分析
S变换
分数阶傅里叶变换
齿轮箱
故障诊断
time-frequency analysis
S transform
fractional Fourier transform
gearbox
fault diagnosis
