Port-Hamiltonian approximation of a nonlinear flow problem. Part I: Space approximation ansatz

09/23/2020
by   Björn Liljegren-Sailer, et al.
0

This paper is on the systematic development of robust and online-efficient approximations for a class of nonlinear partial differential equations on networks. The class includes, e.g., gas pipe network systems described by one-dimensional barotropic Euler equations. All steps necessary in nonlinear model reduction are covered by our analysis. These are the space discretization by conventional methods, the projection-based model order reduction and the complexity reduction of nonlinearities. Special attention is paid to the structure-preservation on all levels. The proposed reduced models are shown to be locally mass conservative, to fulfill energy bounds and to inherit port-Hamiltonian structure. The main ingredients of our analysis are energy-based modeling concepts like the port-Hamiltonian framework and the theory on the Legendre transform, which allow a convenient and general line of argumentation. Moreover, the case of the barotropic Euler equations is examined in more detail and a well-posedness result is proven for their approximation in our framework.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset

Sign in with Google

×

Use your Google Account to sign in to DeepAI

×

Consider DeepAI Pro