Showing 1–2 of 2 results for author: Stahl, N

Certified Reduced Basis Method for the Damped Wave Equations on Networks

Authors: Nadine Stahl, Björn Liljegren-Sailer, Nicole Marheineke

Abstract: In this paper we present a reduced basis method which yields structure-preservation and a tight a posteriori error bound for the simulation of the damped wave equations on networks. The error bound is based on the exponential decay of the energy inside the system and therefore allows for sharp bounds without the need of regularization parameters. The fast convergence of the reduced solution to the… ▽ More In this paper we present a reduced basis method which yields structure-preservation and a tight a posteriori error bound for the simulation of the damped wave equations on networks. The error bound is based on the exponential decay of the energy inside the system and therefore allows for sharp bounds without the need of regularization parameters. The fast convergence of the reduced solution to the truth solution as well as the tightness of the error bound are verified numerically using an academic network as example. △ Less

Submitted 11 April, 2022; originally announced April 2022.

Efficient State Estimation for Gas Pipeline Networks via Low-Rank Approximations

Authors: Nadine Stahl, Nicole Marheineke

Abstract: In this paper we investigate the performance of projection-based low-rank approximations in Kalman filtering. For large-scale gas pipeline networks structure-preserving model order reduction has turned out to be an advantageous way to compute accurate solutions with much less computational effort. For state estimation we propose to combine these low-rank models with Kalman filtering and show the a… ▽ More In this paper we investigate the performance of projection-based low-rank approximations in Kalman filtering. For large-scale gas pipeline networks structure-preserving model order reduction has turned out to be an advantageous way to compute accurate solutions with much less computational effort. For state estimation we propose to combine these low-rank models with Kalman filtering and show the advantages of this procedure to established low-rank Kalman filters in terms of efficiency and quality of the estimate. △ Less

Submitted 4 May, 2021; v1 submitted 31 July, 2020; originally announced July 2020.