中文摘要：

在高噪声及动态条件下，传统算法的跟踪性能受到一定的影响，并可能出现振荡现象。为此，设计一种利用在线估计信号的噪声含量和动态特性来调节跟踪特性的算法，以提高频率跟踪性能。该算法通过克拉克变换将三相电压信号转化成一个相量，并利用有限冲击响应（finiteimpulseresponse，FIR）前置滤波器来抑制噪声影响。利用递归最小二乘滤波器来获得信号的相量估计值，此滤波器能够在动态状态下通过减小遗忘因子来得到快速跟踪性能，而在高噪声下通过增大遗忘因子来提高抗噪声性能。根据相邻两采样点的相角差来计算新的采样问隔。仿真结果表明，该算法能在动态和静态条件下都具有比传统算法更为优越的估计性能。

英文摘要：

Severe noise may distort the response of traditional frequency tracking algorithm and dynamic characteristic of signals may lead tracking frequency to appear oscillation. Therefore, a new frequency tracking algorithm is proposed to improve the performance under variable conditions via adaptively changing its setting according to the estimations of noise density and dynamic characteristics. Three-phase voltage signals are transformed into a phasor by Clarke transformation and then a finite impulse response （FIR） filter is employed to suppress noise, Then, a phasor estimation of supplied signals is attained by applying a recursive least-square filter, which reduces the forgetting factor to get a fast tracking ability under dynamic conditions and increase it to have a good noise rejection performance under severe noise condition. Finally, the sample time interval is calculated via the angle difference between the last two estimations. The simulations＇ results demonstrated that proposed algorithm is superior to the traditional one under both steady conditions and dynamic conditions.