Mixed Algorithm of SINDy and HAVOK for Measure-Based Analysis of Power System with Inverter-based Resources

Abstract: Artificial intelligence and machine learning is enhancing electric grids by offering data analysis tools that can be used to operate the power grid more reliably. However, the complex nonlinear dynamics, particularly when coupled with multi-scale interactions among Inverter-based renewable energy Resources, calls for effective algorithms for power system application. This paper presents affective novel algorithm to detect various nonlinear dynamics, which is built upon: the Sparse Identification of Nonlinear Dynamics method for nonlinear dynamics detection; and Hankel Alternative View of Koopman method for multi-scale decomposition. We show that, by an appropriate integration of the strengths of the two, the mixed algorithm not only can detect the nonlinearity, but also it distinguishes the nonlinearity caused by coupled Inverter-based resources from the more familiar ones caused synchronous generators. This shows that the proposal algorithm can be a promising application of artificial intelligence and machine learning for data measure-based analysis to support operation of power system with integrated renewables.