Data-Driven Modeling of Microgrid Transient Dynamics through Modularized Sparse Identification

Modularized sparse identification (M-SINDy) is developed in this paper for effective data-driven modeling of the nonlinear transient dynamics of microgrid systems. The high penetration of power-electronic interfaces makes microgrids highly susceptible to disturbances, causing severe transients, especially in the islanded mode. The M-SINDy method realizes distributed discovery of nonlinear dynamics by partitioning a higher-order microgrid system into multiple subsystems and introducing pseudo-states to represent the impact of neighboring subsystems. This specific property of the proposed algorithm is found to be very useful while working with re-configurable and scalable microgrids. The governing equations of the subsystems are identified through regression by mapping the nonlinear system's data to a linear system in a large functional space. The effectiveness of the M-SINDy method is tested and validated through numerical examples and comparisons with other existing identification models on a typical islanded microgrid. The method can simultaneously compute the governing equations of different subsystems and is examined to be robust to measurement noises and partial observations. This paper highlights the advances of data science in providing a potent tool for modeling and analyzing higher-order nonlinear microgrid systems. Dynamic discovery of system transients from measurements can be beneficial for designing control strategies that improve the overall microgrid stability and reliability.