摘要
对多分量非平稳信号进行时频分析可以获得信号特征随时间的变化规律,揭示信号的瞬时频率、能量分布和频谱演化。采用传统的时频方法处理多分量信号时,受制于自身算法缺陷,无法对信号进行高分辨率时频表达。逐步衍生出的改进的时频方法,如同步压缩变换(SST)等,在处理噪声干扰下的强时变调频多组分信号时的效果仍然欠佳。调频变换(CT)相对于其他方法,在处理非平稳弱幅值分量时具有较强的适应性与噪声鲁棒性。针对非平稳弱特征信号的时频表示弱这一问题,对CT算法进行了研究。首先,对基于CT衍生出的多分辨率调频变换(MRCT)进行了理论分析;然后,为了进一步提高MRCT的时频分析能力,基于同步提取变换(SET)的原理对MRCT进行了优化,引入了同步提取算子(SEO),并提出了多分辨率同步提取调频变换(MRSECT);最后,为了验证MRSECT在处理复杂振动信号时的有效性,利用MRSECT对数值模拟信号与实验轴承故障数据进行了分析。研究结果表明:MRSECT能够有效提取非平稳信号的时频特性,Renyi熵值在模拟信号处理中提高了25%,其在处理实际轴承故障信号中提高了15.6%。相比于其他时频算法,MRSECT具有更高的时变特征信息提取能力。
The time-frequency analysis of multi-component non-stationary signals can obtain the changing rule of signal characteristics with time,and reveal the instantaneous phase and frequency,energy distribution and spectrum evolution of signals.The traditional time-frequency method for processing multi-component signals was limited by its own algorithmic limitations and could not achieve high-resolution time-frequency representation of the signal.Although improved time-frequency methods such as synchro squeezing transform(SST) were gradually derived,their performance in handling strong time-varying frequency modulated multi-component signals under noise interference was still not satisfactory.Comparing to other methods,chirplet transform(CT) had the adaptability of weak amplitude components and robustness to noise for processing non-stationary signals.Chirplet transform(CT) has strong adaptability and noise robustness in handling non-stationary weak amplitude components compared to other methods.Based on this,firstly a theoretical analysis of the multi-resolution chirplet transform(MRCT) derived from CT was conducted.Then,in order to further improve the time-frequency analysis capability,MRCT was optimized based on the principle of synchro extracting transform(SET),the synchro extracting operator(SEO) was introduced,and the multi-resolution synchro extracting chirplet transform(MRSECT) was proposed.Finally,in order to verify the effectiveness of MRSECT when processing complex vibration signals,MRSECT was used to analyze the numerical simulation signals and experimental bearing fault data.The research results demonstrate that MRSECT can effectively extract the time-frequency characteristics of non-stationary signals,and Renyi entropy increases by 25% in analog signal processing and 15.6% in actual bearing fault signal processing.MRSECT has higher time-varying feature information extraction ability compared to other time-frequency algorithms.
作者
牟令
覃英杰
易灿灿
MOU Ling;QIN Yingjie;YI Cancan(Hubei Energy Group Renewables Development Co.,Ltd.,Wuhan 430069,China;Key Laboratory of Metallurgical Equipment and Control,Ministry of Education,Wuhan University of Science and Technology,Wuhan 430081,China)
出处
《机电工程》
CAS
北大核心
2024年第6期1116-1126,共11页
Journal of Mechanical & Electrical Engineering
基金
国家自然科学基金资助项目(51805382)
湖北省重点研发计划项目(2021BAA194)。
关键词
同步压缩变换
调频变换
多分辨率调频变换
同步提取变换
同步提取算子
多分辨率同步提取调频变换
synchro squeezing transform(SST)
chirplet transform(CT)
multi-resolution chirplet transform(MRCT)
synchro extracting transform(SET)
synchro extracting operator(SEO)
multi-resolution synchro extracting chirplet transform(MRSECT)
