Application of adaptive filters in active power filters

This paper presents a new idea applied to adaptive filters for harmonic detection in shunt active power filters. Improve the speed of convergence of the adaptive filter and reduces the steady-state error are the finals objectives of the strategy. Two cases are presented and discussed, one with a finite impulse response (FIR) adaptive filter with 32 coefficients and another with an adaptive notch filter. The least mean square (LMS) algorithm was used to adjust the coefficients in the both cases. An algorithm based on Coulon oscillator principle was used to determine the load current amplitude and then modify the step-size value. Simulations using Matlab are presented to clarify the algorithm. Also practical implementation is performed using the DSP Texas Instruments TMS320F2812 and the experimental results depicted.