Online Identification of Power Grid Models Via Synchrophasor Data
MetadataShow full item record
Power system dynamics are becoming increasingly complex, as new interconnections are made between systems and more power from renewable resources is integrated into the electric grid. As a result, assessing and maintaining the stability of power systems is becoming more and more challenging. There is a need for better, simpler dynamical models to allow system operators and others to make quick, smart decisions about system operation. Fortunately, the widespread adoption of Phasor Measurement Units (PMUs) has allowed more data to be collected regarding power system operation. This research investigates how to leverage PMU data to produce accurate dynamical models through Bode analysis. Assuming system input is perfectly known and controllable, it is shown that Bode phase and magnitude plots can be constructed empirically by inputting a sum of sine waves into a linear system and observing the output. However, problems arise when this technique is applied to a noisy system, and the resulting Bode plots are shown to be unreliable. Since introducing large sinusoidal oscillations into a power system is undesirable, it is concluded that producing dynamical models through Bode analysis is impractical in real power systems. Instead, future work will investigate model identification methods more resilient to system noise, including recursive parallel, equation error, and output error methods. Once effective modeling methods are identified for linear systems, methods for modeling non-linear systems will be considered. Also, applications of MATLAB’s System Identification Toolbox in deriving these models will be examined.
Wiseman, Benjamin P (2016). Online Identification of Power Grid Models Via Synchrophasor Data. Undergraduate Research Scholars Program. Available electronically from