Power system harmonic parameters estimation using ADALINE-VLLMS algorithm

This paper proposes a combined adaptive linear neural network and variable leaky least mean square (ADALINE-VLLMS) algorithm for harmonic parameters estimation in a distorted power signal. VLLMS method applies a variable leaky adjustment technique for avoiding drifting of the parameter involved in the estimation. Moreover, a variable step-size is proposed to have faster convergence than conventional least mean square (LMS) algorithm. Computer simulations show that the proposed ADALINE-VLLMS method gives better performance than the ADALINE-LMS method.