Sparse Identification-Enabled Data-Driven Modeling for Nonlinear Dynamics of Microgrids

Considering the widespread integration of power-electronic-interfaced distributed energy resources (DERs) in microgrids, the system's inertia is highly reduced. Thus, microgrids are sensitive to disturbances. To get the situational awareness of transient dynamics of microgrids when they face severe disturbances, a data-driven modeling algorithm, Sparse Identification of Non-Linear Dynamics (SINDy), is discussed in this paper. The governing equation of microgrids can be identified through SINDy, which determines the nonlinear dynamics of the system. Numerical examples test and verify the data-driven model based on SINDy can successfully capture the transient dynamics of microgrids. Simulation results have validated the data-driven model is reliable and very close to the actual test system. Therefore, it provides a potent tool to analyze the system's transient dynamics when an accurate system's model is unavailable.