This article analyzes the impact of forced power system oscillations on mode damping estimation. Parametric (Yule-Walker) and non-parametric (Welch) methods for mode estimation are tested in the presence of forced power system oscillations. For mode damping estimation based on non-parametric methods, an application of Half Power Point method is proposed. Performances of the mode estimators are evaluated using both simulated and real synchrophasor data from the Nordic Grid. The presence of forced oscillations poses difficulties to mode damping estimators, these difficulties are identified, illustrated and explained herein.
This article presents a software development toolkit for Wide Area Monitoring Systems. By using this development toolkit, a researcher is able to manipulate synchrophasor data in the LabView environment, which enables fast software prototyping and testing. This toolkit makes full scale testing in real-time easier for researchers, liberating them of complex and time consuming synchrophasor data handling. The toolkit exploits the IEEE C37.118.2-2011 protocol making it independent of any specific equipment and their manufacturers. An application of the development kit is demonstrated in a laboratory environment with a specially designed experimental setup composed of a real-time digital simulator and four phasor measurement units (PMUs).
Existing mode meter algorithms were derived with the assumption that load variations are accurately represented by white noise or an integral of white noise, which may not be satisfied in actual power systems. This paper proposes a mode meter algorithm which relaxes this assumption by explicitly taking into account spectral load characteristics. These characteristics can be either measured or estimated using the inverse of the existing power system model. The method is developed assuming an autoregressive moving average (ARMA) model of the system and incorporating estimated correlations between loads as inputs and synchrophasor measurements as outputs. Performances of the proposed method are compared with the Yule-Walker and N4SID methods using simulated synchrophasor data obtained from the KTH Nordic 32 test system. Finally, the effects of measurement noise on the proposed method are analyzed, as well as the effects of model uncertainty when the power system model is used to determine spectral load characteristics. It is shown that the proposed algorithm increases accuracy in mode estimates when the loads are described with nonwhite noise.