Improved adaptive algorithm of harmonic current detection

Based on the study of disadvantages of conventional fixed step-size LMS (Least Mean Square) algorithm for harmonic current detection, an adaptive LMS algorithm with variable step-size is provided, which adjusts the step-size according to the autocorrelation estimation between e (n) and e (n-D). When the weight is far away from the optimal value, it increases the step-size to accelerate the tracking of time-variable system, while when the weight is near the optimal value, it decreases the step-size to minish the steady state error. By selecting appropriate parameters, the algorithm allows more flexible control of misadjustment and faster convergence speed. Its theoretical approximate expressions are derived. It is easy to be implemented with little additional calculation. The algorithm can effectively adjust the step-size while maintaining the immunity against harmonic current disturbance. Simulation results illustrate its effectiveness.