Dirac and Lagrange algebraic constraints in nonlinear port-Hamiltonian systems (Q2022393): Difference between revisions

Port-Hamiltonian systems provide a systematic approach to the modeling, analysis, simulation, and control of multiphysics systems, possibly on a large scale, see e.g. [\textit{V. Duindam} (ed.) et al., Modeling and control of complex physical systems. The port-Hamiltonian approach. Berlin: Springer (2009; Zbl 1179.93007)] and references therein. In [Syst. Control Lett. 121, 31--37 (2018; Zbl 1408.93019)] the authors gave a generalization of port-Hamiltonian systems obtained by replacing the Hamiltonian function representing the energy storage by a Lagrangian subspace. The above-mentioned article motivated the work by \textit{C. Beattie} et al. [Math. Control Signals Syst. 30, No. 4, Paper No. 17, 27 p. (2018; Zbl 1401.37070)], where the framework of port-Hamiltonian systems has been extended to dynamical systems with linear constraints (descriptor systems, differential algebraic equations) of arbitrary index and with variable constraints in time. On the other hand, \textit{M. Barbero Liñán} et al. [J. Geom. Mech. 11, No. 4, 487--510 (2019; Zbl 1439.37060)] used Morse families to generate Lagrangian submanifolds. Moreover, they generalized the Dirac systems in such a way that the dynamics is defined by those Lagrangian submanifolds of the phase space in the framework of the port-Hamiltonian systems. The present paper extends the above results from the linear to the nonlinear setting. A theoretical framework is established to study the Dirac and Lagrange algebraic restriction equations that arise in port-Hamiltonian (generalized) systems, extending the linear results obtained in [the authors, Syst. Control Lett. 121, 31--37 (2018; Zbl 1408.93019)]. In addition, a definition of nonlinear generalized port-Hamiltonian systems is provided by including energy dissipation and external ports. The authors show how the implicit energy storage relations can be locally represented by a Hamiltonian depending on a part of state variables and a complementary part of co-state variables. Also, by extension of the state space (by Lagrange multipliers), they prove that Dirac algebraic constraints can be converted into Lagrange algebraic constraints, and conversely. Finally the authors claim that this work should be seen as a starting point for further studies on numerical properties of the resulting structured classes of nonlinear differential-algebraic equations systems.