Dynamic Phasor Measurement Unit Test System

This paper presents the plans and progress towards the development of a dynamic phasor measurement unit (PMU) performance test system at NIST. We describe an algorithm for taking time-synchronized samples of single-phase voltage and current power signals and calculating their dynamic parameters, in particular the signal magnitude, phase, frequency, and rate of change of frequency that a PMU reports. PMUs must time stamp their values at periodic Coordinated Universal Time (UTC) markers called the update times. Thus, to provide a reference for PMU testing the sampled data can be fit to a model to define the value of a dynamic parameter at a specific time. The analysis model proposed in this paper assumes that the dynamic magnitude and frequency parameters of the signals are constant over the sampling interval analyzed. This analysis interval is usually the same as the update period or an integer multiple of that period. In the proposed analysis model the dynamic magnitude and frequency parameters are considered a polynomial in time about the update times. The order of the polynomial can be adjusted in a way that meets the needs of the signal being analyzed, yet minimizes the computational effort and sensitivity to noise. We show that when the dynamic variations are analyzed in this way, a single matrix can be used to iteratively converge on a good estimate of the dynamic frequency and magnitude parameters. The polynomial model can be used to generate and analyze test signals. Several test patterns are proposed, which include linearly changing magnitudes or frequencies. As expected, during low voltage tests of the system, the analysis does very well when the generation model matches the analysis model. Several other generation models are also proposed, such as sine waves or damped sine waves. The proposed analysis model is shown to be very accurate in these cases as well.