A Non-Intrusive Method to Inferring Linear Port-Hamiltonian Realizations using Time-Domain Data

TL;DR

This paper presents a non-intrusive approach to identify linear port-Hamiltonian systems from time-domain data by inferring frequency response and constructing realizations, demonstrated through numerical examples and comparisons.

Contribution

It introduces a novel non-intrusive method for constructing port-Hamiltonian systems directly from time-domain data, combining frequency response inference with realization construction.

Findings

Successfully infers frequency response from time-domain data.

Constructs port-Hamiltonian realizations using inferred data.

Outperforms some existing system identification methods.

Abstract

Port-Hamiltonian systems have gained a lot of attention in recent years due to their inherent valuable properties in modeling and control. In this paper, we are interested in constructing linear port-Hamiltonian systems from time-domain input-output data. We discuss a non-intrusive methodology that is comprised of two main ingredients -- (a) inferring frequency response data from time-domain data and (b) constructing an underlying port-Hamiltonian realization using the inferred frequency response data. We illustrate the proposed methodology by means of two numerical examples and also compare it with two other system identification methods to infer the frequency response from the input-output data.