Power System Harmonic Detection Based On Variable Step LMS Algorithm Research

Based on the basic principle of adaptive harmonic detection, this articles make a more in-depth theoretical analysis and research on traditional LMS algorithm. Aiming at the conflict between the convergence rate and steady-state error of fixed step LMS algorithm,this articles proposes a variable step size LMS adaptive algorithm of fast convergence under the load mutation. The algorithm uses an improved coherent averaging theory to adjust the step iterative formula, and introduces momentum to improve the convergence rate of weights, so that the new algorithm has a faster convergence rate and the ability to resist mutation tracking system. And it is applied to the harmonic detection of power system, compared with traditional LMS algorithm, and the experimental simulation show that the new algorithm improves the performance in the harmonic detection of LMS algorithm, good results are obtained.Taking into account the noise interference of actual work environment and the shortcomings of the LMS algorithm, integrating the advantages of the existing variable step size algorithm, this paper proposes a variable step size adaptive filtering algorithm based on the improved hyperbolic function. And the convergence, anti-interference ability and tracking capabilities of the new algorithm are analyzed in theory, finally, the algorithm with the existing variable step size algorithm are compared for filter simulation through the MATLAB software, confirmed the superiority of the proposed variable step LMS algorithm.