Real-time harmonics estimation in power systems using a novel hybrid algorithm

This study presents a new hybrid algorithm to estimate the harmonic parameters of a distorted signal in power systems. The parameters to be estimated are amplitudes and phases of harmonic components according to the voltage/currents samples. The proposed algorithm is based on combination of the recursive least squares (RLS) and iterated extended Kalman filter (IEKF) techniques. The RLS–IEKF algorithm decomposes the problem into linear amplitude estimation and non-linear phase estimation leading to extracting the intended state vector in online mode and intensive noise presence. As well, RLS–IEKF estimates dynamic parameters using tuning factor which controls the impact of measurement on estimation process. Simulation results obtained by MATLAB show the accuracy and speed of convergence in comparison with that of conventional discrete Fourier transform and ensemble Kalman filter. For further validation, the proposed algorithm is implemented by C++ code and is applied to real switching current data. The real-time implementation of RLS–IEKF in a simple laboratory setup using PC/104 computer set and dedicated hardware shows its satisfactory performance for practical power quality and protection cases.