基于希尔伯特—黄变换的次同步振荡检测研究

Subsynchronous Oscillation Detection Based on the Hilbert-Huang Transform Method

针对目前电力系统次同步振荡的辨识局限于线性化方法,提出了一种非线性、非平稳信号的处理方法:希尔伯特—黄变换。首先对振荡信号进行滤噪和时延补偿预处理,然后用黄变换辨识出模态参数,最后与改进的PRONY算法及快速傅里叶变换的辨识结果对比分析。仿真结果表明,经验模态分解可以滤除信号中的噪声干扰,为模态参数的准确分析奠定了基础;并且黄变换方法能有效分解出频率不接近的振荡模态,进而准确的辨识出模态参数,得到振荡模态的时频特性。鉴于此方法会出现频率模态漏分解的情况,在实际工程中可同时使用改进PRONY法、快速傅里叶变换和黄变换以提高次同步振荡辨识的准确度。

Subsynchronous oscillation in power system is conventionally detected by linear method,so Hilbert-Huang transform method is proposed as a solution to nonlinear and non-stationary signals.Firstly,the subsynchronous oscillation signal is preproccessed through filtering and time-delay compensation.Then oscillation mode parameters are identified by Hilbert-Huang transform.Finally,the method is compared with modified PRONY algorithm and fast Fourier transform so as to prove its availability.Simulation results show that empirical mode decomposition can filter the noise efficiently,reinforcing the basis of accurate analysis of mode parameters.Hilbert-Huang transform is effective in identifying oscillation modes and their parameters,and presents oscillation modes＇ frequency-time property.In case of the missing detection of frequency modes,it is suggested to use the improved PRONY algorithm,fast Fourier transform and Hilbert-Huang transform simultaneously in order to enhance the accuracy of subsynchronous oscillation detection.