Efficient state estimation for gas pipeline networks via low-rank approximations

Abstract

In this paper we investigate the performance of projection-based low-rank approximations in Kalman filtering. For large-scale gas pipeline networks structurepreserving 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.