基于变步长LMS自适应陷波算法的电气信号频率测量

Estimation of electrical signal frequency based on adaptive notch filter with variable step-size LMS algorithm

提出一种基于二阶无限冲击响应(IIR)变步长自适应数字陷波滤波器的电气信号频率跟踪测量算法。论述了陷波滤波器能够滤除信号中特定的频率,而其他的频率成分不受影响的原理;最小均方LMS(LeastMeanSquare)自适应陷波算法是将被谐波和随机噪声污染的电气信号通过基波陷波器,根据陷波器输出误差采用变步长因子的递推LMS自适应修正陷波器参数和跟踪频率的变化。实例中,给出了给定变步长迭代公式的常数以计算出频率,并采用频率稳定时的测量、频率波动时的测量、电机运行频率测量的3种仿真结果表明所提出的频率跟踪测量算法效果良好。

An approach for estimating the frequency of electrical signal disturbed by harmonics and noise is presented,which is based on adaptive second-order ⅡR（Infinite Impulse Response） notch filter with variable step-size LMS（Least Mean Square） algorithm. The working principle of notch filter is discussed in detail,which removes specific frequencies from input signal without influencing other frequency components. The notch filter coefficients are adapted according to the error of its outputs by a recursive LMS algorithm with variable step-size. In an example,with the given constants of variable step-size iterative formula,the frequency is calculated. Three simulative estimations of stable frequency,varying frequency and motor operating frequency are presented to demonstrate the effec-tiveness of the proposed algorithm.