基于TLS算法三相系统的谐波检测*

将TLS、LMS和约束反Hebbian自学习算法统一起来，指出LMS算法和约束反Hebbian自学习算法是两个特例。研究了TLS自适应学习方法在谐波检测中的应用，并将这种方法推广到三相系统中。所研究的检测方法同时适用于单相和三相对称非对称负载电流的谐波检测，可对谐波电流、无功电流以及负序电流进行分别或综合检测。仿真实验证明，在跟随性能和误差性能方面，单相TLS算法要优于单相LMS和约束反Hebbian算法，而三相TLS检测方法的性能要明显优于按单相的TLS检测方法。

Harmonic detection for three-phase circuit based on TLS algorithm

Based on the Newton learning algorithm,this paper presented a new algorithm for TLS estimation and its application in harmonic detection. It was shown that the LMS algorithm and constrained anti-Hebbian algorithm were two examples of the proposed algorithm. The dynamic and error performances of the TLS algorithm were better than that of LMS and anti-Hebbian algorithms. And the APF based on TLS algorithm could be used under the conditions of three-phase unbalanced and nonlinear loads, could be applicable to detect harmonic, reactive and negative sequence currents respectively or the sum of ones. At the same time, the smaller error than that in single-phase can be obtained.