Conference article

DAE Solvers for Large-Scale Hybrid Models

Erik Henningsson
Dassault Systèmes AB, Lund, Sweden

Hans Olsson
Dassault Systèmes AB, Lund, Sweden

Luigi Vanfretti
Rensselaer Polytechnic Institute, Troy, NY, USA

Download articlehttp://dx.doi.org/10.3384/ecp19157491

Published in: Proceedings of the 13th International Modelica Conference, Regensburg, Germany, March 4–6, 2019

Linköping Electronic Conference Proceedings 157:50, p. 12

Show more +

Published: 2019-02-01

ISBN: 978-91-7685-122-7

ISSN: 1650-3686 (print), 1650-3740 (online)

Abstract

We present a strategy for DAE mode simulations of large-scale Modelica models with state events. DAE solvers can be orders of magnitudes faster than traditional ODE solvers when simulating models with large algebraic loops. Such loops are common in, for example, power grid models. Central for our DAE mode approach is the accurate and efficient treatment of state events. Adapting, extending, and optimizing results known in the literature to the Modelica context resulted in a DAE mode implementation first released in Dymola 2019 and 3DEXPERIENCE 2019x. The implementation is verified by efficiency experiments featuring OpenIPSL power grid models. The run times for these models are competitive with domain-specific, state-of-the-art simulation tools.

Keywords

DAE mode, hybrid model, state event, largescale, Modelica, power grid model

References

No references available

Citations in Crossref