Improving the convergence time of adaptive notch filters to harmonic detection

Abstract

This paper presents the application of adaptive notch filters for harmonic detection in shunt active power filters (SAPF). The orthogonal input signals to the adaptive notch filter are generated using the Clarke transformation of the load currents in such way that the Phased Looked Loop (PLL) is not necessary any more. The least mean square (LMS) and recursive least square (RLS) algorithms were used to adjust the coefficients of the adaptive notch filter. The final objective of the described strategy is to improve the speed of convergence of the adaptive notch filter, reducing the steady-state error and the transient response. All these features are very important to real time applications in active power filters. Simulations using Matlab/Simulink are presented to clarify the algorithm. Also, the implementation of the technique in a shunt active power filter is performed using the DSP Texas Instruments TMS320F2812, and the experimental results depicted.