Design of Robust Temperature Control for Low Temperature Zones in District Heating Grids
Authors
Hvims, Jacob Ørsnes ; Rindom, Ida Agerbo
Term
4. term
Education
Publication year
2022
Submitted on
2022-05-30
Pages
116
Abstract
Fjernvarmen ønsker at sænke fremløbstemperaturen for bedre at integrere vedvarende energi, men i lavtemperaturzoner kan stor afstand mellem blandingspunkt og temperatursensor give en betydelig transportforsinkelse, som forringer reguleringen. Denne afhandling undersøger, hvordan en robust temperaturregulator kan udformes til at mindske indflydelsen af denne forsinkelse og forbedre temperatursporing på trods af forstyrrelser samt sæson- og parametervariationer. Tre variationer af en Smith Predictor-regulator foreslås for at kompensere for den gennemstrømningsafhængige forsinkelse, og en PI-regulator designes til at sikre robust stabilitet ved hjælp af en usikkerhedsmodel, der indbefatter variationer i rørlængder og -diametre for sommer- og vinterscenarier. De foreslåede strukturer sammenlignes i simuleringer med en klassisk regulator under temperaturforstyrrelser. Resultaterne viser, at Smith Predictor-varianterne ikke forbedrer temperatursporing, fordi kompensationsleddet ikke kan forudsige forstyrrelserne, og at der derfor kræves en anden kontrolløsning.
District heating networks aim to lower supply temperatures to better integrate renewable energy, but in low-temperature zones a long distance between the mixing point and the temperature sensor introduces transport delay that degrades control. This thesis investigates how to design a robust temperature controller that reduces the impact of this delay and improves temperature tracking despite disturbances and seasonal and parameter variations. Three variations of a Smith Predictor control scheme are proposed to compensate the flow-dependent delay, and a PI controller is designed to ensure robust stability using an uncertain plant model that includes variations in pipe lengths and diameters for summer and winter scenarios. The proposed schemes are evaluated in simulation against a classical control solution under temperature disturbances. The results show that the Smith Predictor variants do not improve tracking because the compensation term cannot predict disturbances, indicating that another control approach is needed.
[This summary has been generated with the help of AI directly from the project (PDF)]
Documents