针对傅里叶分解方法存在过度分解、运算时间长等问题,提出了一种基于循环频谱包络的经验傅里叶分解(CEEFD)算法,并将该算法运用到滚动轴承故障诊断中。首先,对信号进行了快速傅里叶变换(FFT),获得了信号的频谱,对傅里叶频谱进行了循环包...针对傅里叶分解方法存在过度分解、运算时间长等问题,提出了一种基于循环频谱包络的经验傅里叶分解(CEEFD)算法,并将该算法运用到滚动轴承故障诊断中。首先,对信号进行了快速傅里叶变换(FFT),获得了信号的频谱,对傅里叶频谱进行了循环包络,得到了包络曲线,减少了无用极值点的个数,抑制了噪声对分量的干扰;然后,采用改进的局部最大最小值(local max min)分割技术,对频谱包络曲线进行了频带分割;最后,构建了零相位滤波器,采用逆快速傅里叶变换(IFFT)对每个频带进行了信号重构,得到了若干个瞬时频率且具有物理意义的单分量信号;通过对仿真信号和滚动轴承实测信号的分析,并将其与经验模态分解(EMD)、经验小波变换(EWT)、傅里叶分解方法(FDM)、变分模态分解(VMD)和经验傅里叶分解(EFD)进行了实验对比验证。研究结果表明:采用CEEFD方法获得的单分量包含了更准确的故障特征信息,验证了CEEFD方法的有效性,CEEFD方法可用于轴承的故障诊断;相对于上述方法,CEEFD方法具有更高的准确精度和更强的抗噪声干扰能力。展开更多
In nanoparticle sizing using the ultrafast image-based dynamic light scattering (UIDLS)method,larger impurities and dark noise from the complementary metal-oxide-semiconductor (CMOS)detector affect measurement accurac...In nanoparticle sizing using the ultrafast image-based dynamic light scattering (UIDLS)method,larger impurities and dark noise from the complementary metal-oxide-semiconductor (CMOS)detector affect measurement accuracy.To solve this problem,a two-dimensional self-adapting fast Fourier transform (2D-SAFFT)algorithm is proposed for UIDLS.Dynamic light scattering images of nanoparticles are processed using 2D fast Fourier transforms,and a high-frequency threshold and a low-frequency threshold are then set using the self-adapting algorithm to eliminate the effects of the dark noise of the CMOS detector and the impurities.The signals caused by the dark noise of the CMOS detector and the impurities are cut off using the high-frequency threshold and the low-frequency threshold.The signals without the high- and low-frequency components are then processed again using an inverse Fourier transform to obtain new images without the dark noise and impurities signals.The mean diameters of the measured nanoparticles can be obtained from images obtained using UIDLS.Five standard latex nanoparticles (46,100, 203,508,994nm)and commercial nanoparticles (antimony-doped tin oxide,indium tin oxide,TWEEN- 80,nano-Fe,and nano-Al2O3)were measured using this new method.Results show that 2D-SAFFT can effectively eliminate the effects of dark noise from the CMOS detector and the impurities.展开更多
A non-invasive detection method for the status analysis of cell culture is presented based on digital holography technology.Lensless Fourier transform digital holography (LFTDH) configuration is developed for living...A non-invasive detection method for the status analysis of cell culture is presented based on digital holography technology.Lensless Fourier transform digital holography (LFTDH) configuration is developed for living cell imaging without prestaining.Complex amplitude information is reconstructed by a single inverse fast Fourier transform,and the phase aberration is corrected through the two-step phase subtraction method.The image segmentation is then applied to the automatic evaluation of confiuency.Finally,the cervical cancer cell TZMbl is employed for experimental validation,and the results demonstrate that LFTDH imaging with the corresponding image post-processing can provide an automatic and non-invasive approach for monitoring living cell culture.展开更多
文摘针对傅里叶分解方法存在过度分解、运算时间长等问题,提出了一种基于循环频谱包络的经验傅里叶分解(CEEFD)算法,并将该算法运用到滚动轴承故障诊断中。首先,对信号进行了快速傅里叶变换(FFT),获得了信号的频谱,对傅里叶频谱进行了循环包络,得到了包络曲线,减少了无用极值点的个数,抑制了噪声对分量的干扰;然后,采用改进的局部最大最小值(local max min)分割技术,对频谱包络曲线进行了频带分割;最后,构建了零相位滤波器,采用逆快速傅里叶变换(IFFT)对每个频带进行了信号重构,得到了若干个瞬时频率且具有物理意义的单分量信号;通过对仿真信号和滚动轴承实测信号的分析,并将其与经验模态分解(EMD)、经验小波变换(EWT)、傅里叶分解方法(FDM)、变分模态分解(VMD)和经验傅里叶分解(EFD)进行了实验对比验证。研究结果表明:采用CEEFD方法获得的单分量包含了更准确的故障特征信息,验证了CEEFD方法的有效性,CEEFD方法可用于轴承的故障诊断;相对于上述方法,CEEFD方法具有更高的准确精度和更强的抗噪声干扰能力。
基金National Natural Science Foundation of China (Grant No.51573093).
文摘In nanoparticle sizing using the ultrafast image-based dynamic light scattering (UIDLS)method,larger impurities and dark noise from the complementary metal-oxide-semiconductor (CMOS)detector affect measurement accuracy.To solve this problem,a two-dimensional self-adapting fast Fourier transform (2D-SAFFT)algorithm is proposed for UIDLS.Dynamic light scattering images of nanoparticles are processed using 2D fast Fourier transforms,and a high-frequency threshold and a low-frequency threshold are then set using the self-adapting algorithm to eliminate the effects of the dark noise of the CMOS detector and the impurities.The signals caused by the dark noise of the CMOS detector and the impurities are cut off using the high-frequency threshold and the low-frequency threshold.The signals without the high- and low-frequency components are then processed again using an inverse Fourier transform to obtain new images without the dark noise and impurities signals.The mean diameters of the measured nanoparticles can be obtained from images obtained using UIDLS.Five standard latex nanoparticles (46,100, 203,508,994nm)and commercial nanoparticles (antimony-doped tin oxide,indium tin oxide,TWEEN- 80,nano-Fe,and nano-Al2O3)were measured using this new method.Results show that 2D-SAFFT can effectively eliminate the effects of dark noise from the CMOS detector and the impurities.
基金supported by the National Natural Science Foundation of China (No.61077004)the Science Foundation of Education Commission of Beijing (No.KZ200910005001)and the Innovative Talent and Team Development for Serving Beijing
文摘A non-invasive detection method for the status analysis of cell culture is presented based on digital holography technology.Lensless Fourier transform digital holography (LFTDH) configuration is developed for living cell imaging without prestaining.Complex amplitude information is reconstructed by a single inverse fast Fourier transform,and the phase aberration is corrected through the two-step phase subtraction method.The image segmentation is then applied to the automatic evaluation of confiuency.Finally,the cervical cancer cell TZMbl is employed for experimental validation,and the results demonstrate that LFTDH imaging with the corresponding image post-processing can provide an automatic and non-invasive approach for monitoring living cell culture.