An Adaptive Time-Efficient Technique for Harmonic Estimation of Nonstationary Signals

Model-based parametric techniques offer many advantages over conventional discrete Fourier transform-based methods for harmonic/interharmonic estimation. However, high computational requirements restrict their applications to offline analysis purpose. In this paper, an adaptive technique based on estimation of signal parameters via rotational invariance technique (ESPRIT) is proposed that optimizes the accuracy and computation time for harmonic/interharmonic estimation of stationary as well as nonstationary power supply signals. This method first estimates the order of the model (number of sinusoids present in the distorted power supply signal) and then adjusts the autocorrelation matrix dimension based on reconstruction error. The performance of the proposed method is validated on the time-varying simulated signal, measured synthetic signal, and actual voltage signal of a distribution system supplying electric arc welding load. The comparison of the results with the short-time Fourier transform and the sliding window ESPRIT techniques shows that the proposed approach considerably reduces the computational time of high-resolution ESPRIT method along with better accuracy.