Condition monitoring of degradation parameters using dynamic mode decomposition

Abstract

The parameter estimation of the nonlinear system demands complex estimation algorithms. Recently data-driven modeling such as Dynamic Mode Decomposition (DMD) and Dynamic Mode Decomposition with Control (DMDc), has gained popularity in the field of parameter estimation and system identification of complex non-linear systems. In this paper, we have applied the DMDc technique for system identification and parameter estimation of the DC motor and DC-DC buck converter with and without knowledge of the input matrix. Further, this helps in condition monitoring and predictive maintenance of different passive components in the buck converter, such as capacitors and resistors. The results conclude that the DMD estimates those parameters within a tolerance limit of 2.2% with a minimum number of data required to capture the dynamics of the system. © 2023 IEEE.