Abstract
A comprehensive procedure to identify a potable water system by data-driven means is proposed. The methodology focuses in a first step on topology identification. Herein, the sensor data is used to establish a directed graph of the underlying system by the means of a coherence-based algorithm and heuristics within a multi-stage strategy. In a second step, the pipe lengths of the water system are identified with the usage of dead time estimates from regularized finite impulse response models. The whole procedure is successfully applied to a simulation model. It is shown that the topology can be correctly identified. The pipe lengths can be determined with sufficient accuracy if heat transfer between supply pipes and return pipes can be neglected. The proposed method can be used to identify the topology and pipe lengths of potable water systems.
Zusammenfassung
Es wird ein neues datengetriebenes Verfahren zur Identifikation von Trinkwasserinstallationen vorgestellt. Die Methodik konzentriert sich in einem ersten Schritt auf die Topologie-Identifikation. Es werden Sensordaten verwendet, um einen gerichteten Graphen eines zugrunde liegenden Systems zu erstellen. Dafür wird eine mehrstufige Strategie erstellt, welche einen auf der Berechnung von Kohärenzfunktionen basierenden Algorithmus und Heuristiken nutzt. In einem zweiten Schritt werden die Rohrlängen des Wassersystems unter Verwendung von Totzeitschätzungen mittels regularisierten Finite Impulse Response Modellen identifiziert. Das vorgestellte Verfahren wird erfolgreich auf ein Simulationsmodell angewendet. Es wird gezeigt, dass die Topologie korrekt identifiziert werden kann. Die Rohrlängen können ausreichend genau bestimmt werden, wenn die Wärmeübertragung zwischen Vor- und Rücklauf vernachlässigt werden kann. Die vorgestellte Methode kann genutzt werden, um die Topologie und Rohrlängen einer Trinkwasserinstallation zu identifizieren.
About the authors
Timm J. Peter graduated with a Master of Science degree from Universität Siegen in 2018. After finishing his masters thesis about regularized FIR models he joined the working group Automatic Control – Mechatronics of Prof. Nelles as a research assistant. His research topics focus on new techniques for linear and nonlinear system identification.
Max Schüssler is a research assistant with the working group Automatic Control – Mechatronics of Prof. Nelles. In his work he focuses on machine learning perspectives for nonlinear system identification and adjacent research fields. He graduated with a Master of Science degree from Universität Siegen in 2018.
Daniel Rüschen received the M. Sc. degree in Electrical Engineering, Information Technology and Computer Engineering as well as the Dr.-Ing. degree in Control Engineering from RWTH Aachen University. After his time as a Research Associate at the Philips Chair for Medical Information Technology at RWTH Aachen University, he joined the Advanced Development Department of Viega where he focuses on energy efficient operation of drinking water installations using control engineering methods.
Oliver Nelles is Professor at the University of Siegen in the Department of Mechanical Engineering and chair of Automatic Control – Mechatronics. He received his doctor’s degree in 1999 at the Technical University of Darmstadt. His key research topics are nonlinear system identification, dynamics representations, design of experiments, metamodeling and local model networks.
-
Author contributions: Timm J. Peter and Max Schüssler contributed equally.
References
1. Deutsche Energie-Agentur. Der dena-Gebäudereport: Statistiken und Analysen zur Energieeffizienz im Gebäudebestand. Deutsche Energie-Agentur, 2016.Search in Google Scholar
2. Deutsche Energie-Agentur. Der dena-Gebäudereport: Statistiken und Analysen zur Energieeffizienz im Gebäudebestand. Deutsche Energie-Agentur, 2019.Search in Google Scholar
3. Nir Friedman and Daphne Koller. Being bayesian about network structure. A bayesian approach to structure discovery in bayesian networks. Machine learning, 50(1-2):95–125, 2003.10.1023/A:1020249912095Search in Google Scholar
4. Trevor Hastie, Robert Tibshirani and Jerome H. Friedman. The elements of statistical learning. Springer series in statistics. Springer, New York, NY, 2nd edition, 2009.10.1007/978-0-387-84858-7Search in Google Scholar
5. Giacomo Innocenti and Donatello Materassi. A modeling approach to multivariate analysis and clusterization theory. Journal of physics A: mathematical and theoretical, 41(20), 2008.10.1088/1751-8113/41/20/205101Search in Google Scholar
6. Rosario Mantegna and H. Eugene Stanley. An Introduction to Econophysics: Correlations and Complexity in Finance. 08 2007.Search in Google Scholar
7. Donatello Materassi and Giacomo Innocenti. Unveiling the connectivity structure of financial networks via high-frequency analysis. Physica A: Statistical Mechanics and its Applications, 388(18):3866–3878, 2009.10.1016/j.physa.2009.06.003Search in Google Scholar
8. Donatello Materassi and Giacomo Innocenti. Topological identification in networks of dynamical systems. IEEE Transactions on Automatic Control, 55(8):1860–1871, 2010.10.1109/CDC.2008.4739317Search in Google Scholar
9. Charles D. Michener and Robert R. Sokal. A quantitative approach to a problem in classification. Evolution, 11(2):130–162, 1957.10.1111/j.1558-5646.1957.tb02884.xSearch in Google Scholar
10. Oliver Nelles. Nonlinear System Identification: From Classical Approaches to Neural Networks and Fuzzy Models. Springer, Berlin, 2001.10.1007/978-3-662-04323-3Search in Google Scholar
11. Klaus Stierstadt. Thermodynamik: von der Mikrophysik zur Makrophysik. Springer, 2010.10.1007/978-3-642-05098-5Search in Google Scholar
12. Marc Timme. Revealing network connectivity from response dynamics. Physical review letters, 98(22):224101, 2007.10.1103/PhysRevLett.98.224101Search in Google Scholar PubMed
© 2021 Walter de Gruyter GmbH, Berlin/Boston
