This paper discusses the joint use of control configuration selection and adaptive model predictive control for the reconfiguration and improvement of a control system for the heat generation and distribution in a district heating system. The traditional PID-type multi-loop controller configuration is assessed from its feasibility and potential to be improved, on the basis of automatically generated models. The studied test case is the district heating system in Luleå, Sweden.
It is found that a multivariable controller for a subsystem of the complete system could largely improve the performance of the system and thereby addressing confirmed oscillations in the system. An adaptive model predictive controller is implemented in the realistic simulator OPTi-Sim for the district heating system in Luleå, Simulation results confirm that the oscillations which are present in the simulation with the current control scheme can be largely reduced through the use of the proposed controlled.
ISBN för värdpublikation: 978-3-907144-00-8, 978-1-7281-1314-2