Conference article

Model-order selection of output-error models – BSM1 as case study

Christian Wallin
ABB AB, Power Generation, Västerås, Sweden / School of Business, Society and Engineering, Mälardalen University, Västerås, Sweden

Jesús Zambrano
School of Business, Society and Engineering, Mälardalen University, Västerås, Sweden

Download articlehttp://dx.doi.org/10.3384/ecp18153243

Published in: Proceedings of The 59th Conference on Simulation and Modelling (SIMS 59), 26-28 September 2018, Oslo Metropolitan University, Norway

Linköping Electronic Conference Proceedings 153:34, p. 243-247

Show more +

Published: 2018-11-19

ISBN: 978-91-7685-494-5

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

Abstract

Output-Error (OE) System Identification is used to estimate the nonlinear behavior of an activated sludge process (ASP) in aWastewater Treatment Plant (WWTP). The aim is to identify dynamic models to reproduce the effect of different plant dynamics. How the dissolved oxygen concentration of the aerobic tank affect the effluent ammonia concentration and how the internal recirculation affect the nitrate concentration of the anoxic tank is studied. The best fit of the model is estimated by varying the model order through a trial-and-error approach. Three different scenarios are investigated: one Single-Input-Single-Output (SISO) and two Multiple-Input-Multiple-Output (MIMO) structures. In the SISO scenario only the oxygen to the effluent ammonia dynamics is investigated. Then for both the MIMO scenarios the internal recirculation to nitrate concentration dynamics in the anoxic tank is included and in the last scenario the influent flow rate is also included. The approach is evaluated using the Benchmark Simulation Model no.1 (BSM1).

Keywords

Benchmark Simulation Model No. 1, Model Predictive Control, Output Error Model, System Identification. Strict

References

No references available

Citations in Crossref