A Novel Technique for Fundamental and Harmonic Parameter Estimation Using Nonergodic $S$ -Transform

Abstract

Modern power systems produce nonstationary harmonics which bring harm to the electric equipment. Real-time and accurate parameter estimation is the key step of the harmonic elimination. This paper proposes a novel technique of the parameter estimation for fundamental and harmonic using nonergodic $S$ -transform (NEST), which can effectively eliminates the estimated parameter error caused by nonsynchronous sampling. First, the signal is decomposed into the fundamental and harmonics NEST vectors (NESTVs) by NEST, and the amplitudes are detected accurately by extracting the magnitude of NESTVs, which are optimized by the Gaussian window width coefficient of the NEST kernel function. Second, the overlapping and abandoning method is used to remove the end effect of NESTVs so as to obtain the effective and complete data for the parameter estimation. Then, the estimated instantaneous frequency and phase are obtained using the linear phase characteristic of NESTVs. Finally, simulations and experiments prove that the proposed method can not only estimate the stationary signal parameters accurately but also trace the nonstationary signal parameters in the real time under the spectrum leakage condition.