摘要
数字信号的频谱分析中,DFT得到的频谱只能粗略确定实验信号各谐波频率,振幅和相位,单频谱谐波在其频率的某一邻域内的细化幅值频谱和相位频谱具有显著的特征,分析比较表明,单频率谐波细化频谱与矩形窗的频谱极为相似,依此为基础,可以判定密集频率信号,进而通过待定谐波参数,选择合适的参数区间和步长组合循环计算,并用矩形窗频谱近似单频率谐波细化频谱的办法,则可以还原校正密集频率的谐波参数,校正精度略低于细化频谱对单频率谐波的计算结果,该方法可以较好的进行参数多变的密集频率频谱分析,密集的频率越多,计算量也会更大。
In the process of digital signal processing, the spectrum obtained with DFT can only roughly determine frequency, amplitude and phase of each harmonic component of a test signal. For a single-frequency harmonic, its amplitude zoom spectrum and phase spectrum have distinct features in a neighbor field of its frequency. Here, through analysis and comparison, a single-frequency harmonic zoom spectrum and a rectangular window spectrum were very similar. Based on this law, a digital signal including several close frequency harmonics could be identified. Furthermore, appropriate parameter range and step size were selected for loop calculations to determine harmonic parameters. A rectangular window spectrum was used to approximate a signle-frequency zoom spectrum. Therefore, harmonic parameters could be identified and corrected from a digital signal with close frequencies. Calculation precision was slightly lower than that of a single-frequency harmonic zoom spectrum calculation. It was shown that the proposed method can conduct better the spectral analysis with close frequencies in cases of varying parameters; the more the close frequencies, the more the calculations.
出处
《振动与冲击》
EI
CSCD
北大核心
2013年第2期171-174,181,共5页
Journal of Vibration and Shock
关键词
密集频谱
离散傅里叶变换
矩形窗频谱
细化频谱
close frequencies
discrete Fourier transformation （DFT）
rectangular window spectrum
zoom spec-trum
