HARMONIC COMPONENTS ESTIMATION IN POWER SYSTEM USING BACTERIAL FORAGING OPTIMIZATION ALGORITHM AND STOCHASTIC GRADIENT ALGORITHM WITH VARIABLE FORGETTING FACTOR

In this paper, a hybrid configuration algorithm called stochastic gradient method with variable forgetting factor (SGVFF) is proposed to better estimate unknown parameters in a power system such as amplitude and phase of harmonics using variable forgetting factor following the bacterial foraging optimization algorithm (BFO). It must be mentioned that harmonic estimation is a nonlinear problem and using linear optimization algorithms for solving this problem reduces the convergence speed. Thus, BFO algorithm is used for initial estimation. In this paper, first, using little information and by applying BFO algorithm in an off-line procedure initial value for SGVFF algorithm is achieved and then SGVFF algorithm is gained in an on-line procedure. In the hybrid algorithm applied in this paper, amplitudes and phases are estimated simultaneously. Simulation results indicate that the proposed method has faster convergence speed, better performance and higher accuracy in a noisy system in comparison with recursive least squares variable forgetting factors algorithm (RLSVFF). This proves the superiority of the proposed method.