guest ::
%0 Thesis %A Kümpel, Alexander %T Adaptive agentenbasierte modellprädiktive Regelung für Gebäudeenergiesysteme; 1. Auflage %V 140 %I RWTH Aachen University %V Dissertation %C Aachen %M RWTH-2025-04636 %@ 978-3-948234-54-6 %B E.On Energy Research Center %P 1 Online-Ressource : Illustrationen %D 2025 %Z Druckausgabe: 2025. - Auch veröffentlicht auf dem Publikationsserver der RWTH Aachen University %Z Dissertation, RWTH Aachen University, 2024 %X Reducing energy consumption and greenhouse gas emissions in the building sector requires energy efficient operation of buildings. One promising method is model predictive control (MPC), which uses a mathematical model to determine the optimal mode of operation. The advantages of MPC are predictive operation, exploitation of flexibility, multi-variable control, and consideration of multiple targets. The development of a suitable model, tuning, and implementation of MPC are expensive and inhibit its widespread use in the building sector. The objective of this work is to develop a self-adjusting model predictive control through an adaptive and modular approach, which can be applied to various building energy systems with low configuration efforts, to reduce the barriers of MPC for practical use. The basic idea of the approach is to divide the energy system into recurrent subsystems. A hierarchical agent-based approach is used to control the subsystems, where agents control the subsystems using adaptive MPC. Adaptive MPC allows the agents to be transferred to subsystems of the same type with little configuration effort. For efficient operation of the overall system, a coordinator is used to solve a high-level optimization problem. The optimization problem is modular, analogous to the agents, and based on a heat flux-based approach. The specific cost function and model equations of the optimization problem are determined by the individual agents and given to the coordinator. The coordinator determines the setpoints for the individual subsystems by solving the high-level optimization problem and passes them to the agents. A simulation model based on a real building energy system is developed to evaluate the developed control. For a realistic evaluation, the model is calibrated and validated with measurement data. Further, the control of the building energy system is implemented as a reference. In comparison to the reference control, the developed approach leads to a higher energy efficiency and an improved quality of control for the control of various subsystems. Furthermore, the agents are applicable to subsystems of the same type. The agent-based control of the overall energy system results in cost savings up to 59 %F PUB:(DE-HGF)11 ; PUB:(DE-HGF)3 %9 Dissertation / PhD ThesisBook %R 10.18154/RWTH-2025-04636 %U https://publications.rwth-aachen.de/record/1011646