Data-Driven Feedback Linearization Control of Distributed Energy Resources Using Sparse Regression

A complex physics-based modeling procedure and the uncertainty and confidentiality of internal parameters of distributed energy resources (DERs) motivate system identification tools for control purposes in smart grids. This paper develops a framework for data-driven nonlinear modeling and control of DERs using sparse identification of nonlinear dynamics (SINDy). Using the proposed data-driven model for closed-loop control, we demonstrate the effectiveness of a model-free design in stability analysis of DERs in smart grids. Feedback linearization control of DERs was chosen over conventional vector control in this research due to its superior capability of accounting for DER nonlinearities and weak AC grid integration. Compared with existing physics-based designs that heavily rely on knowing the detailed system dynamics or uninterpretable data-driven designs that rely on large historical data, the proposed model-free DER identification and control framework can accurately capture the dynamics of the DERs based on available measurements and provide guaranteed performance for black-start, weak AC grid integration, microgrid integration, and stability analysis. Real-time and offline simulations in addition to a detailed eigenvalue analysis are conducted to compare the effectiveness of the proposed data-driven approach with physics-based controllers.