利用改进数位演算法和经验模态分解的电力系统低频振荡信号瞬时参数的提取

An Improved Prototype Algorithm Based on Empirical Mode Decomposition and Its Application in Extracting Instantaneous Parameters of Low Frequency Oscillation

为分析电力系统低频振荡的动态特性,需要提取低频振荡信号的瞬时参数,而电力系统低频振荡信号为非线性时变复合信号,因此参数提取前需要对其进行振荡模态分解。首先结合最小平方误差法提出了改进数位演算方法,以提高信号参数的提取精度;其次应用经验模态分解法对低频振荡信号进行模态分解,确定有效振荡模态;最后根据提取参数可分析电力系统低频振荡的动态特性。算例结果验证了该方法的有效性。

To analyze dynamic characteristics of low frequency oscillations occurred in power systems, it is necessary to extract instantaneous parameters of low frequency oscillation, however the signal of low frequency oscillation is a nonlinear time-varying compound signal, so the empirical mode decomposition （EMD） of low frequency oscillation should be performed before the extraction of parameters. Firstly, combining with least square error an improved prototype algorithm is proposed to improve the extraction accuracy; then by use of EMD the mode decomposition of low frequency oscillation signal is carried out to determine effective oscillation model; finally, according to the extracted parameters the dynamic characteristics of low frequency oscillation occurred in power system can be analyzed. The effectiveness of the proposed method is verified by simulation results.