Noise Amplitude in Ambient PMU Data and its Impact on Load Models Identification

Abstract

A current trend in load modeling topic is to take advantage of ambient data from Phasor Measurement Units (PMU) to estimate the parameters of load models. In this context, the estimation algorithms or methodologies that are proposed or investigated need to be evaluated in a controlled environment, where, among other things, synthetic PMU measurements obtained from simulations are used. These synthetic measurements require the addition of noise to be like the real ones. The problem found in the literature is the large difference in noise magnitudes used by the authors in their research. These magnitudes in several cases are inconsistent with each other and even seem to be exaggerated. It is for this reason that the present work determines the noise contained in the ambient data reported by PMU. The reliability of the results of this work is based, among other things, on the use of real PMU measurements, located in two different countries, with diverse reporting rates, and located at high, medium, and low voltage. Moreover, this work quantifies the impact that noise has on load modeling with ambient PMU data. In conclusion, the main results of this work are two. The first one covers the noise magnitudes contained in ambient PMU data. The second one demonstrates that noise has a significant and negative impact on load modeling.