基于蝴蝶优化算法改进的参数化重采样时频变换  

Parameterized Resampling Time-Frequency Transform Improved by Butterfly Optimization Algorithm

作　　者：郑直 王叶仑 贾晓龙 ZHENG Zhi;WANG Yelun;JIA Xiaolong(College of Mechanical Engineering,North China University of Science and Technology,Tangshan Hebei 063210,China)

机构地区：[1]华北理工大学机械工程学院,河北唐山063210

出　　处：《机床与液压》2025年第7期24-30,共7页Machine Tool & Hydraulics

基　　金：河北省自然科学基金项目(E2022209086);唐山市科技创新团队培养计划项目(21130208D);河北省科技重大专项项目(22282203Z)。

摘　　要：参数化重采样时频变换(PRTFT)在处理多模态信号时,其窗口长度L选取存在经验性和主观性,导致模态分量的能量分散和时频轨迹模糊等问题,而背景噪声干扰又加剧了上述问题。基于此,引入蝴蝶优化算法(BOA)对PRTFT进行优化选取最优窗口长度L,进而获取多模态信号的各个模态分量的较高能量集中度和较高时频分辨率;采用改进反演定理对多模态信号进行消噪处理,进一步获得更高的能量集中度和清晰时频轨迹。通过滚动轴承内圈故障实验验证分析可知,经BOA优化后的PRTFT可有效地提升信号能量集中度和时频分辨率,且较麻雀搜索算法、星鸦优化算法具有优越性。When the multimodal signal is processed by parameterized resampling time-frequency transform(PRTFT),the selection of window lengthLis empirical and subjective,which leads to the energy dispersion of modal components and fuzzy time-frequency trajectory and other issues,and the background noise interference exacerbates the above problems.Based on the above,the butterfly optimization algorithm(BOA)was introduced to optimize PRTFT to select the optimal window length L for obtaining the higher energy concentration and time-frequency resolution of each modal component of the multi-modal signal.Then,the multi-mode signal was denoised by improved inversion theorem to obtain higher energy concentration and clear time-frequency trajectory.The experimental results show that the BOA optimized PRTFT can effectively improve the signal energy concentration and time-frequency resolution,and they are more effective and superior to the sparrow search algorithm and the nutcracker optimization algorithm.

关 键 词：参数化重采样时频变换 蝴蝶优化算法 时频表示 

分 类 号：TH133.3[机械工程—机械制造及自动化]