基于Kaiser窗双谱线插值FFT的谐波分析方法

为进一步减少加窗插值FFT的频谱泄漏和栅栏效应,提出了基于Kaiser窗的双谱线插值FFT的电力谐波分析方法,运用多项式拟合求出了实用的插值修正算式,推导了信号基波与各次谐波频率、幅值、初相角的计算式。仿真结果表明,Kaiser窗函数设计实现灵活、抑制频谱泄漏效果好,基于Kaiser窗的双谱线插值FFT方法能有效克服基波频率波动与白噪声对谐波分析的影响,在非整数周期截断条件下,对含21次谐波信号的频率计算相对误差仅为1.4×10%,幅值计算相对误差≤0.002%,初相位计算相对误差≤0.0001%,三相谐波电能计量应用实践证明了该方法的正确性。     

基于加窗插值FFT的加速度传感器信号处理方法

针对快速傅里叶变换在加速度传感器动态特性分析时产生的频谱泄漏等问题,提出一种基于加窗插值FFT的加速度传感器信号处理方法。该方法选取精度较高的汉宁窗和三次样条插值函数对信号进行处理,抑制频谱泄漏等带来的干扰,并通过经验传递函数估计方法得到高精度的频率响应估计。实验结果表明,该方法能够有效地提高加速度传感器频率响应函数的估计精度。     

非同步采样下电力系统相量测量修正算法

非同步采样条件下,利用傅里叶变换进行相量分析将出现由频谱泄漏造成的误差。针对工频交流正弦信号中基波电压偏移且采样频率恒定而引起的非同步采样误差,推导了非同步采样下傅里叶变换结果表达式,建立了离散傅里叶变换相角与幅值的误差模型,得到相角和幅值误差与频率偏移量和采样初始相位角的函数关系,提出了基于基波正序分量法的相角与幅值修正算法。所提算法不需要对信号施加特殊的窗函数。仿真和实测结果表明,在基波频率波动剧烈、高斯白噪声存在、非同步采样程度较大的情况下,利用较少的计算量便可准确地获得信号的相角和幅值,适用于电力系统相量的高精度快速测量。     

道路模拟试验随机谱加窗重构方法研究

运用损伤编辑理论可对道路模拟试验随机谱进行编制,达到删除小载荷实现加速试验的目的。但较少文献涉及如何光滑连接并重构删除小载荷后的保留载荷片段这一关键技术问题。提出一种基于信号加窗重叠的道路模拟试验随机谱连接与重构方法。对比了几种常用窗函数对实测随机谱的连接重构效果,发现汉宁窗不仅可有效解决两段信号的光滑连接重构问题,同时也不会引入极值载荷。以某整车试验场实测载荷为数据基础,结合时域损伤编辑法运用加汉宁窗方法构建了试验加速谱,从幅值域和频率域对加窗谱的有效性和合理性进行了理论验证,发现加窗谱完整的保留了未加窗谱的损伤值且两者的功率谱密度曲线高度重合。最后,应用该加速谱建立了整车道路模拟试验,试验响应谱与目标谱达到了高度一致,验证了该方法的有效性。     

基于Kaiser窗相位差校正的电力谐波分析与应用

采用基本窗函数和广义余弦窗函数对信号加权可减少频谱泄漏和栅栏效应的影响,但其效果受到窗函数固定旁瓣性能的制约.Kaiser窗可定义一组可调的窗函数,自由选择主瓣与旁瓣衰减之间的比重,因此能全面地反映主瓣与旁瓣衰减之间的交换关系.对信号非同步采样和非整周期截断时的频谱进行了分析,讨论了Kaiser窗的主、旁瓣特性,提出了基于Kaiser窗相位差校正的电力谐波分析方法,推导了信号基波与各次谐波频率、幅值、初相角的计算式.仿真结果表明:Kaiser窗函数设计实现灵活、抑制频谱泄漏效果好,Kaiser窗相位差校正算法克服了基波频率波动对谐波分析的影响,对含21次谐波信号的基波频率计算相对误差仅为1.0×10-7%,幅值计算相对误差≤0.000 6%,初相位计算相对误差≤0.008%,三相谐波分析与谐波电能计量应用实践证明了该方法的正确性.     

Nuttall窗加权谐波分析算法及其在电能计量中的应用

采用FFT进行电网信号谐波参数分析时存在频谱泄漏,准确度较低.研究提出了一种基于Nuttall窗函数的频谱相位差校正算法,分析了Nuttall窗函数的时域、频域特性,建立了频谱相位差校正算法的流程,推导了信号谐波参数计算式,给出了算法在三相谐波电能计量中的实现方法.采用Nuttall窗对电压、电流信号进行加权,利用Nuttall窗旁瓣电平低、旁瓣衰减快的特点可有效减少频谱泄漏:运用频谱相位差校正方法进行信号频谱校正与分析,不必求解高次方程即可实现电网信号谐波参数分析.仿真与实际运行结果表明,非同步采样情况下,本文算法可以准确地实现电网信号谐波分析,实现电能的准确计量.     

基于Nuttall窗频谱校正的介质损耗因数测量

采用旁瓣峰值电平小且旁瓣渐近衰减速率大的窗函数对信号加权可减轻非同步采样和数据截断对介质损耗因数tanδ测量的影响。本文分析了Nuttall窗的旁瓣特性,提出了基于Nuttall窗双谱线插值FFT的tanδ测量方法,运用多项式拟合求出了实用的插值修正公式,推导了信号初相角及tanδ的计算式,并介绍了基于该算法的测量系统的硬件实现方案。仿真和试验结果表明,Nuttall窗函数抑制频谱泄漏效果好,基于Nuttall窗的双谱线插值FFT方法克服了谐波干扰、基波频率波动及白噪声对tanδ测量的影响,且设计实现灵活,测量结果精确、稳定,可用于介质损耗因数tanδ的离线测量与在线监测。     

基于四项余弦窗三谱线插值FFT的谐波检测方法

快速傅里叶变换(fast Fourier transform,FFT)在非同步采样情况下存在较大的误差,为了减小非同步采样对FFT的影响,通过选择旁瓣性能良好的4项5阶Nuttall窗结合插值FFT进行电力谐波检测。推导了实用的三谱线插值修正公式,利用谐波频点附近的三根离散频谱的幅值确定谐波谱线的准确位置,进而得到谐波幅值、频率及相位的修正公式。仿真结果验证了该算法在非同步采样时,与加Hanning窗和Blackman-Harris窗插值FFT相比有更高的分析准确度,同时,在加相同窗函数情况下,与双谱线插值修正算法相比也具有更高的准确度。仿真结果还验证了测量数据含有白噪声时,算法仍然具有较高的准确度。基于该算法的检测仪器的实测结果也验证了算法的有效性。     

飞秒激光频域干涉测距相位信号分析与处理

针对飞秒激光测距频谱干涉条纹信号,分析了被测距离精度对于相位信息采集与处理要求,详细介绍了光谱分辨干涉测距的数据处理过程。根据测距信号干涉谱的数学关系,推导了被测距离求解公式。采用快速傅里叶变换和坐标旋转数字计算法进行光谱分辨干涉信号处理,经过时域信号的傅里叶变换、信号相位解包裹、加窗函数时域信号滤波等环节,分别对飞秒激光测距频谱干涉条纹信号中每个频率的相位值进行处理和变换。系统仿真中,离散频谱分析的幅值最大误差从36.3%降低到15.1%。经过比对实验,该法能实现飞秒激光测距信号有效处理。     

消除负频率影响的频谱校正

采用比值法校正低频成分的参数精度比较低，其原因是模型中忽略了负频率成分的贡献。为解决这一问题，基于包含负频率贡献的频谱公式，建立新的校正方案，给出矩形窗和汉宁窗对应的校正公式，采用数值仿真对校正公式进行考核。结果表明，提出的校正公式正确，尤其是汉宁窗的频率定位精度、幅值相对精度和相位精度几乎到双精度运算的精度下限。     

MIMO随机加正弦振动试验J-阻抗复合控制算法

以多输入多输出(multiple-input multiple-output,简称MIMO)随机加正弦混合试验控制系统中信号分离方法、控制策略及试验为研究对象,提出了针对混合型振动试验的分离控制策略,采用相关积分法分离混合信号中的正弦信号与随机信号成分。推导了MIMO线性随机振动系统的反馈控制方程,提出了控制目标针对性更强的Jacobi控制算法。分析了多点正弦振动试验中存在的过修正问题,基于压缩因子修正提出阻抗控制方法,并将其与Jacobi算法共同应用于混合试验系统。以悬臂梁为试验对象,开展J-阻抗复合控制算法在两输入两输出随机加正弦混合试验系统中应用验证。结果表明,算法可以有效地抑制超标谱线,并将结果控制在容差带内,为混合振动试验控制提供有力的理论支撑。     

任意概率分布随机振动激励源的设计与实现

随机振动台是航空航天产品重要的研制、鉴定、可靠性及生产试验设备,提高振动模拟试验的逼真性和振动试验的效果都需要产生非高斯分布的随机激励信号.本文对具有给定功率谱密度分布的非高斯随机振动激励信号的生成技术进行研究与分析,使数字式随机振动控制系统可以同时具有频谱和幅度分布的控制能力.给出了采用数字信号处理技术高精度和快速地产生任意概率分布的随机振动时域激励信号的方法.采用这种方法得到的宽带随机信号的信噪比可达到70 dB,可用于精密随机振动台的驱动控制信号,准确地实现被试验产品的模拟实验.计算机仿真和实验结果也证实了该方法的正确性和有效性.     

Iterative design of DSP cosine windows by placing spectrum zeros,and some new classes of DSP cosine windows

Windows are used in digital signal processing in order to reduce leakage in spectral analysis and to estimate signal parameters in non-coherent sampling. This paper deals with design theory and presents an algorithm that enables the design of cosine windows with spectra fulfilling various prescribed properties (belonging to various window classes) and window order up to 10. The design is based on iterative window spectrum zero placing. The proposed algorithm was implemented in Matlab, but other mathematical tools can also be used. In addition to the design of all three classes of Rife and Vincent windows and of flattop windows, this method also enables the design of windows with different properties, corresponding to five newly-defined window classes. Both symmetrical and periodic (DFT) windows can be designed. The use of windowing for a range of signal parameter estimation is mentioned and referenced, the principle of the window design method is described, and definitions of the new window classes and examples of designed windows are presented. A graphical user interface of the window design described in this paper in VBA-Excel is accessible free of charge on the www site http://measure.feld.cvut.cz/en/DCW.     

基于泄漏对消的电力谐波相角高精度估计算法

采用快速傅里叶变换(FFT)进行电力系统谐波分析时,受频谱泄漏和栅栏效应的影响,相角估计误差往往较大.为减小这类误差,提出一种基于频谱泄漏对消的谐波相角高精度估计算法,结合余弦组合窗时域和频域特性,综合考虑长、短范围频谱泄漏的影响,推导了基于余弦组合窗的插值FFT谐波相角计算式.该算法首先运用基于余弦组合窗的插值FFT方法获得频率偏差估计值,然后采用离散频谱泄漏对消方法进行信号谐波相角估计,实现简单,准确度高.非同步采样情况下,采用典型余弦组合窗(Hanning、Blackman、Blackman-Harris、Rife-Vincent(I)、Nuttall等)进行信号谐波相角估计的仿真实验与应用验证了本文算法的准确性和有效性.     

Better narrowband spectral measurements

Spectrum analysis based on traditional FFT techniques is a widely available, commonly understood and general purpose tool for use with mechanical systems. In applications such as system identification, modal analysis and rotating machinery signature analysis, relevant information content in signals is greatest near narrowband phenomena in the frequency domain. Effort should be focused on reducing bias error in these regions in preference to reducing variance error. Time aliasing methods of spectrum analysis can result in superior bias error reduction, producing an accurately sampled spectral estimate and permitting window function characteristics to be decoupled from the DFT resolution, or frequency line spacing. Windows created with time aliasing can be optimized by adjusting the amount of time or frequency localization, according to the type of spectrum being measured. Assuming conventional criteria, time aliased windows can achieve better mainlobe and sidelobe characteristics than other windows. An innovative approach, appropriate with continuous spectra, is to use the window as a discrete sampling function in frequency to eliminate bias error inherent in the effective convolution between window and signal transform. Both approaches have been successfully implemented and results are presented using time aliased spectral estimates showing more accurate measurements of narrowband phenomena when used in applications of rotating machinery vibration and system identification.     

一种MIMO非高斯振动的逆多步预测法

为了实现多输入多输出(multi-input-multi-output,简称MIMO)非高斯振动试验控制,提出了一种MIMO非高斯振动逆多步预测法。首先,该方法基于广义高斯分布构造零记忆非线性变换,并使用该变换生成非高斯参考信号;其次,推导系统的逆多步预测模型,将参考信号作为该模型的输入,生成试验所需的驱动信号;然后,在控制过程中分别根据控制点的功率谱密度和峭度与相应控制目标之间的误差,对参考非高斯信号进行调整,并利用逆多步预测模型对驱动信号进行更新,以实现非高斯振动控制;最后,进行三轴振动台试验。试验结果表明,该方法对功率谱与峭度都取得了良好的控制效果,验证了其可行性。     

电液振动台加速度随机振动控制

随机振动模拟试验通过时域波形再现非平稳的振动环境,广泛应用于各类产品的可靠性检测。电液振动台是大型结构件进行模拟试验的关键设备,在随机振动环境下,传统三状态控制器无法同时消除电液振动台干扰力与模型不确定性的影响,以单轴电液振动台作为控制对象,利用滑模控制策略,抑制电液振动台系统在随机振动环境中不确定干扰力的影响,利用自适应逆控制器补偿模型不确定性问题,实现振动台加速度随机振动控制。通过MATLAB/Simulink对该复合控制策略进行随机振动仿真分析,验证控制策略的有效性。     

A new technique based on traditional wavelet transform used in NVH application of internal combustion engine

The sound and vibration signal of internal combustion engine (ICE) is typically unsteady and very complex. Recently the wavelet transform, we see, is being widely used as a time–frequency representation. A new technique that is based on traditional wavelet transform and has many advantages over traditional spectrograms is proposed in this study. The performance of the new technique is validated by applying it to a numerical simulation and test signals. It is shown that the new technique is suited for complicated signals that contain multiple impacts and/or dynamic changes in time and frequency domain.     

A study on the determination of optimal reference spectrum for random vibration control in environmental vibration test

For random vibration control in environmental vibration test it is desired that the spectra at the specimen attachment points coincide with the specified reference spectrum. However, this goal is seldom achieved because of the dynamic characteristics of shaker, fixture, and specimen and of the interactions between them in the test frequency ranges. This paper proposes a method for the determination of the optimal reference spectrum which minimizes the spectral deviations between the specified reference spectrum and spectra at the specimen attachment points. The least square method is used for this purpose. This method requires only the measured FRFs at the control point and specimen attachment points in pre-vibration test and is shown through theoretical and experimental results to be effective.     

WINDOW AND OVERLAP PROCESSING EFFECTS ON POWER ESTIMATES FROM SPECTRA

Fast Fourier transform (FFT) spectral processing is based on the assumption of stationary ergodic data. In engineering practice, the assumption is often violated and non-stationary data processed. Data windows are commonly used to reduce leakage by decreasing the signal amplitudes near the boundaries of the discrete samples. With certain combinations of non-stationary signals and windows, the temporal weighting may attenuate important signal characteristics to adversely affect any subsequent processing. In other words, the window artificially reduces a significant section of the time signal. Consequently, spectra and overall power estimated from the affected samples are unreliable. FFT processing can be particularly problematic when the signal consists of randomly occurring transients superimposed on a more continuous signal. Overlap processing is commonly used in this situation to improve the estimates. However, the results again depend on the temporal character of the signal in relation to the window weighting. A worst-case scenario, a short-duration half sine pulse, is used to illustrate the relationship between overlap percentage and resulting power estimates. The power estimates are shown to depend on the temporal behaviour of the square of overlapped window segments. An analysis shows that power estimates may be obtained to within 0.27 dB for the following windows and overlap combinations: rectangular (0% overlap), Hanning (62.5% overlap), Hamming (60.35% overlap) and flat-top (82.25% overlap).