Active Power Measurements - An Overview and a Comparison of DSP Algorithms by Noncoherent Sampling

Abstract

This paper presents an overview of algorithms for one-phase active power estimation using digital signal processing in the time domain and in the frequency domain, and compares the properties of these algorithms for a sinusoidal test signal. The comparison involves not only algorithms that have already been published, but also a new algorithm. Additional information concerning some known algorithms is also included. We present the results of computer simulations in MATLAB and measurement results gained by means of computer plug-in boards, both multiplexed and using simultaneous signal sampling. The use of new cosine windows with a recently published iterative algorithm is also included, and the influence of additive noise in the test signal is evaluated.

Synchronization of Sampling-Based Measuring Systesm

“…Interpolation algorithms [15]- [17] lead to good approximations of the amplitude and frequency, however, they require to be optimized according to the weighted function applied to the measured signal [18], [19]. Moreover, even if the performances can be very competitive [8], [20], [21], the expected accuracy is still below that can be obtained from a simple DFT analysis of a data set obtained by coherent sampling.…”

“…Such ADCs are particularly well suited to metrological applications such as impedance calibration by sampling techniques [1]- [3], development of Josephson-locked sine-wave synthesizer [4], [5], and electrical power measurements [6]- [8].…”

On the Use of Multi-Harmonic Least-Squares Fitting for THD Estimation in Power Quality Analysis

“…However, the estimation of the FFT requires a further strain on the acquisition system because it requires synchronous acquisition to ensure that any frequency variations do not cause spectral leakage and errors in the estimation of the harmonic amplitudes and on the final THD value [8]. In [9] an efficient alternative to the use of coherent sampling was proposed for active power estimation, based on windowing. Time-frequency algorithms such as the GaborWigner Transform (GWT) have also been proposed for power quality assessment [10].…”

Estimation of the Amplitude Square Using the Interpolated Discrete Fourier Transform

“…Therefore, these coefficients can be written in two parts (7)(8)(9): the larger term due to the short-range spectrum leakage of the component investigated m , and bias   i  due to the long-range leakage of the spectral image of this component ( Fig. 1) and also due to the long-range leakage contributions from other components in the multicomponent signal.…”

“…Assuming non-coherency in the sampling process, spectral granularity and leakage may adversely affect the accuracy of the estimation process. To cope with such issues, various methods have been devised for the accurate estimation of tone amplitudes and their squares [9], and among them two main methods are frequently used: the energy-based method estimates the power in each spectral component by evaluating the total energy falling inside a band including the window main lobe [8,10], and the interpolated DFT estimates the amplitude of the spectral lines of interest using two or more neighboring DFT coefficients, starting from those centered in each spectrum's local maximum [11][12][13]. While the energy-based method is a more intuitive technique and only needs generic window specifications, the interpolated WDFT requires more calculations and a thorough knowledge of spectral window behavior but performs with a reduced systematic bias error as shown in this paper.…”