Stepwise Commissioning: of a Steam Boiler
Author
Johansen, Simon Vestergaard
Term
4. term
Education
Publication year
2015
Submitted on
2015-06-03
Pages
93
Abstract
Dette projekt undersøger, hvordan man bedst starter og indkører avancerede regulatorer til industrielle systemer med flere input og output (MIMO). Målet er at gøre idriftsættelsen mere ligetil ved først at køre med et sæt enklere enkelt-sløjferegulatorer (SISO) og derefter trinvis skifte over til en fuld MIMO-regulator, når anlægget er i drift. Til dette formål er der udviklet en trinvis idriftsættelsesstrategi baseret på Youla–Kucera-parametriseringen, en matematisk metode der gør det muligt at blande mellem stabiliserende regulatorer på en kontrolleret måde og samtidig sikre nominel stabilitet (stabilitet for den antagne model). De eneste ekstra krav er en model af anlægget (en plantemodel) samt mindst én yderligere stabil/stabiliserende regulator. For at afprøve strategien i praksis er der gennemført et casestudie i samarbejde med Grundfos på en dampkedel hos Danish Crown i Sønderborg. Opgaven var at designe en fødevandsregulator til at styre vandstanden i kedlen, hvilket gør det til et MISO-problem (flere input, ét output). En model er overtaget fra et tidligere studenterprojekt, og parametre er estimeret ud fra de samme data. Med denne model er der designet både en PI-regulator (proportional–integral) og en optimal MISO-regulator, som indgår i den trinvise strategi. Regulatoren er implementeret på en styreplatform og testet på den rigtige dampkedel. Den gradvise overgang mellem de to regulatorer lignede tæt den teoretiske, hvilket gør strategien lovende i praksis.
This project examines how to start up and commission advanced controllers for industrial systems with multiple inputs and outputs (MIMO). The goal is to make commissioning more straightforward by first running a set of simpler single-loop controllers (SISO) and then gradually transitioning to a full MIMO controller once the plant is operating. To enable this, a stepwise commissioning strategy based on the Youla–Kucera parameterization was developed. This mathematical framework allows engineers to blend between stabilizing controllers in a controlled way while guaranteeing nominal stability (stability for the assumed model). The only additional requirements are a model of the plant and at least one additional stable, stabilizing controller. To test the strategy in practice, a case study was conducted with Grundfos on a steam boiler at Danish Crown in Sønderborg. The task was to design a feedwater controller to regulate the boiler water level, which makes it a multiple-input single-output (MISO) problem. A model from a previous student project was adopted, and parameter identification was carried out using the same data. Based on this model, both a proportional–integral (PI) controller and an optimal MISO controller were designed and used in the stepwise strategy. The controller was implemented on a control platform and tested on the real boiler. The gradual transition between the two controllers closely matched the theoretical predictions, indicating that the strategy is practical and promising.
[This abstract was generated with the help of AI]
Documents