The Periodic Characteristics of Harmonic Measurement Errors with the Initial Sampling Time

Abstract

The stability and consistency of harmonic amplitude measurement is very important for power quality analysis and assessment. Some random factors may influence the measurement results in the real world, for example the initial sampling time. In this paper, we put this view a little forward, and prove that the amplitude errors with DFT method change periodically with the initial-sampling time, and DFT method “may” give relatively worse results, with the window which has better characteristics. To mitigate this random influence and to achieve more stable measurement precision, the well known shifting window average DFT, which is based on expectation calculation, is adopted. The effectiveness of this method and its recursive mode has been tested on several simulated signals and measured signals generated by Electrical Power Standards Fluke 6100A.