Digital twins in engineering dynamics: variational equation identification, feedback control design and their rapid update.

Digital twins, as a rising concept in many industry sectors, have tremendous potential for engineering dynamics. Their practical applications, however, are hindered by the lack of near real-time system identification and control design methods. This paper establishes and implements a framework of digital twins in engineering dynamics based on the latest advances in the field of data-driven equation identification. It consists of a preprocessing step and a circular step; the former includes an initial identification of the digital twin (i.e., an integral variational equation) from discrete data captured from a practical physical twin and an initial design of feedback control to intervene in the physical and digital twins simultaneously; the latter includes the detection of system variation and the rapid update of the digital twin and control strategy. Two typical examples, i.e., a Duffing oscillator and a slender straight beam with time-varying force along its axis, are adopted to illustrate the application and efficacy of the established framework, with the different control objectives of suppressing deterministic/stochastic responses and improving system stability. [ABSTRACT FROM AUTHOR]