Real-time harmonic estimation using a novel hybrid technique for embedded system implementation

摘要

In this paper, an analytical technique for efficient estimation of harmonics generated by power systems is presented. A robust linear estimator known as H∞ filter is applied to extract the amplitudes of harmonics. As well, phases are estimated using a nonlinear algorithm, namely, cubature Kalman filter. Both methods are based upon space modeling of the studied signals leading to improved numerical stability. The proposed H∞-cubature Kalman filter algorithm has the capability of estimating power system time varying waveforms even in highly noisy measurements. Synthetic and real signals are used for testing the proposed algorithm, and accuracy of the estimation is reported by statistical indices. Simulations in MATLAB are compared with unscented Kalman filter to represent the quality of estimation. Real-time implementation of the proposed algorithm is conducted using C++ programming code in hardware- in-the-loop laboratory. The setup consists of an embedded computer set and dedicated hardware. Results show satisfactory performance of the proposed method for practical application in power quality and protection cases.