A study on data-driven identification and representation of nonlinear dynamical systems with a physics-integrated deep learning approach: Koopman operators and nonlinear normal modes (Q6105292)

The spectral theory of the Koopman operators is essential in nonlinear dynamical systems and control theory. The main result of this paper is the physics-integrated deep learning-based approach for identification of the eigenfunction of Koopman operators and the nonlinear modal transformation function of nonlinear normal modes. It was concluded that approach could fail in case of highly nonlinear behavior where the response is quasi-periodic or non-regular. The efficiency of proposed approach is demonstrated using the numerical experiments on the nonlinear Duffing systems of various dimensions.