Affiliation:
1. Chair of Computer Science VI: Artificial Intelligence and Knowledge Systems, Julius-Maximilians-Universität Würzburg, Am Hubland, 97074 Würzburg, Germany
2. ENER-IQ GmbH, Leightonstraße 3, 97074 Würzburg, Germany
Abstract
In a time where sustainability and CO2 efficiency are of ever-increasing importance, heating systems deserve special considerations. Despite well-functioning hardware, inefficiencies may arise when controller parameters are not well chosen. While monitoring systems could help to identify such issues, they lack improvement suggestions. One possible solution would be the use of digital twins; however, critical values such as the water consumption of the residents can often not be acquired for accurate models. To address this issue, coarse models can be employed to generate quantitative predictions, which can then be interpreted qualitatively to assess “better or worse” system behavior. In this paper, we present a simulation and calibration framework as well as a preprocessing module. These components can be run locally or deployed as containerized microservices and are easy to interface with existing data acquisition infrastructure. We evaluate the two main operating modes, namely automatic model calibration, using measured data, and the optimization of controller parameters. Our results show that using a coarse model of a real heating system and data augmentation through preprocessing, it is possible to achieve an acceptable fit of partially incomplete measured data, and that the calibrated model can subsequently be used to perform an optimization of the controller parameters in regard to the simulated boiler gas consumption.
Funder
German Federal Ministry for Economic Affairs and Climate Action
Subject
Electrical and Electronic Engineering,Biochemistry,Instrumentation,Atomic and Molecular Physics, and Optics,Analytical Chemistry
Reference47 articles.
1. Calibrating detailed building energy simulation programs with measured data—Part I: General methodology (RP-1051);Reddy;Hvac&R Res.,2007
2. Automatic calibration of a distributed catchment model;Eckhardt;J. Hydrol.,2001
3. Tun, W., Wong, K.W.J., and Ling, S.H. (2023). Advancing Fault Detection in HVAC Systems: Unifying Gramian Angular Field and 2D Deep Convolutional Neural Networks for Enhanced Performance. Sensors, 23.
4. Matetić, I., Štajduhar, I., Wolf, I., and Ljubic, S. (2023). Improving the Efficiency of Fan Coil Units in Hotel Buildings through Deep-Learning-Based Fault Detection. Sensors, 23.
5. Kosonen, I. (2007, January 1–2). Microscopic freeway simulation with automatic calibration. Proceedings of the International Symposium on Intelligence Techniques in Computer Games and Simulations, Ritsumeikan University, Shiga, Japan.