空间滤波转速遥测系统的信号处理及其实现方法

发布时间：2018-10-26 12:55

【摘要】：转速测量与工业生产、国防建设和科研等密切相关。基于空间滤波技术的转速遥测系统具有结构简单、非接触、可以遥测的特点,有良好的应用前景。在空间滤波转速遥测系统中,从差分空间滤波传感器输出的准正弦信号中提取其中心频率,再用中心频率计算出转速。从准正弦信号中瞬时准确地提取出中心频率是该系统能准确遥测瞬时转速的关键。然而,空间滤波信号是窄带非平稳的,其幅度和相位都有一定的随机性,这使得准确提取空间滤波信号的瞬时中心频率很困难。为解决该难题,根据准正弦信号的特点和小波变换的时频局部性,提出了基于复小波变换的中心频率提取方法(WT方法)。在该方法中,先根据中心频率的预测值构造出有不同中心频率的复小波,用复小波族对准正弦信号进行复小波变换并取其模值,以最大模值对应的小波中心频率作为信号的中心频率。为简单有效地实现复小波变换,用高斯函数的m阶和m+1阶导数作为一对正交小波,用该对正交小波分别作为实部和虚部而构成复母小波;用高斯函数任意阶导数的频谱经离散傅里叶逆变换得到复小波的时域序列,以高效地构建出复小波。用仿真实验和空间滤波器信号,验证了所提出的中心频率提取方法及其性能。在仿真实验中用WT方法提取已知频率正弦信号的频率,通过与快速傅里叶变换方法(FFT方法)、过零点周期方法(CZ方法)提取的正弦信号频率比较,得出了WT方法能够用于提取信号频率的结论。在用实际空间滤波信号进行的实验中,分别采用对称实小波(包括奇、偶对称实小波)和复小波进行了中心频率的提取,通过和FFT方法、CZ方法的实验结果比较,验证了采用实小波变换不能提取出空间滤波信号中心频率,也验证了WT方法比常用的FFT方法、CZ方法瞬时性强、精度高的结论。这样,本文有效的解决了瞬时、准确的提取准正弦信号中心频率的难题。
[Abstract]:Rotational speed measurement is closely related to industrial production, national defense construction and scientific research. The rotational speed telemetry system based on spatial filtering technology has the characteristics of simple structure, non-contact, and can be telemetry, so it has a good application prospect. In the space filtering rotational speed telemetry system, the center frequency is extracted from the quasi sinusoidal signal output from the differential spatial filter sensor, and the rotational speed is calculated by the center frequency. Instantaneous and accurate extraction of center frequency from quasi-sinusoidal signal is the key to accurate telemetry of instantaneous rotational speed. However, the spatial filtering signal is nonstationary in narrow band, and its amplitude and phase are both random, which makes it difficult to accurately extract the instantaneous center frequency of the spatial filter signal. In order to solve this problem, according to the characteristics of quasi-sinusoidal signal and the time-frequency localization of wavelet transform, a method of center frequency extraction based on complex wavelet transform (WT method) is proposed. In this method, the complex wavelet with different central frequency is constructed according to the predicted value of the center frequency, and the complex wavelet transform is performed on the sinusoidal signal by using the complex wavelet family and the modulus of the complex wavelet transform is obtained. The wavelet center frequency corresponding to the maximum modulus value is taken as the central frequency of the signal. In order to realize complex wavelet transform simply and effectively, the m order and m 1 derivative of Gao Si function are used as a pair of orthogonal wavelets, and this pair of orthogonal wavelets are used as real part and imaginary part respectively to form complex mother wavelets. The time-domain sequence of complex wavelet is obtained by inverse discrete Fourier transform with the spectrum of arbitrary derivative of Gao Si function, and the complex wavelet is constructed efficiently. Simulation experiments and spatial filter signals are used to verify the proposed method and its performance. In the simulation experiment, the frequency of the known frequency sinusoidal signal is extracted by WT method. The frequency of the sinusoidal signal is compared with that extracted by the fast Fourier transform method (FFT method) and the zero-crossing period method (CZ method). It is concluded that the WT method can be used to extract the signal frequency. In the experiments using real spatial filtering signals, symmetric real wavelets (including odd and even symmetric real wavelets) and complex wavelets are used to extract the center frequencies respectively. The experimental results are compared with the experimental results of FFT method and CZ method. It is verified that the real wavelet transform can not extract the center frequency of the spatial filter signal, and the conclusion that WT method is more instantaneous and accurate than the usual FFT method is verified. In this way, the problem of instantaneous and accurate extraction of center frequency of quasi-sinusoidal signal is solved effectively in this paper.
【学位授予单位】：重庆理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN911.7

【相似文献】