Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems: A Case Study in Denmark

Frederik Wagner Madsen; Theis Bank; Henrik Alexander Nissen Søndergaard; Lasse Kappel Mortensen; Hamid Reza Shaker

doi:10.1007/978-3-031-48652-4_18

Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems: A Case Study in Denmark

Frederik Wagner Madsen, Theis Bank, Henrik Alexander Nissen Søndergaard^*, Lasse Kappel Mortensen, Hamid Reza Shaker

^*Corresponding author for this work

University of Southern Denmark

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

3 Downloads (Pure)

Abstract

Faults in district heating systems (DHS) cause suboptimal operating conditions, which increase energy losses. As DHSs are critical infrastructure for many households in Denmark, these faults should be detected and corrected quickly. A novel model-based fault detection and diagnosis framework has been applied to detect and prioritise faults. The framework uses a bound for normal operation based on the residuals between historical sensor data and simulated properties in a digital twin of the DHS. The faults detected are prioritised based on the fault probability calculated using the Chernoff bound method. A case study on a Danish DHS has proven that the framework can produce a prioritised list of faults that maintenance crews can use to target faults with the highest probability. Furthermore, the digital twin allowed for fault location investigation, which could correlate different faults in the DHS. The framework has the potential for real-time fault detection and diagnosis. However, more precise digital twins need to be developed.

Original language	English
Title of host publication	Energy Informatics : Third Energy Informatics Academy Conference, EI.A 2023, Campinas, Brazil, December 6–8, 2023, Proceedings, Part II
Editors	Bo Nørregaard, Luiz Carlos Pereira da Silva, Zheng Ma
Volume	2
Publisher	Springer
Publication date	2024
Pages	277–291
ISBN (Print)	978-3-031-48651-7
ISBN (Electronic)	978-3-031-48652-4
DOIs	https://doi.org/10.1007/978-3-031-48652-4_18
Publication status	Published - 2024
Event	Energy Informatics.Academy Conference 2023 - Unicamp campus, São Paulo , Brazil Duration: 6. Dec 2023 → 8. Dec 2023 Conference number: 3 https://www.energyinformatics.academy/eia-2023-conference

Conference

Conference	Energy Informatics.Academy Conference 2023
Number	3
Location	Unicamp campus
Country/Territory	Brazil
City	São Paulo
Period	06/12/2023 → 08/12/2023
Internet address	https://www.energyinformatics.academy/eia-2023-conference

Series	Lecture Notes in Computer Science
Volume	14468
ISSN	0302-9743

Keywords

fault detection and diagnosis
district heating systems
digital twin
Chernoff bound

Documents & Links

10.1007/978-3-031-48652-4_18

Open Access VersionAccepted manuscript, 346 KB

Cite this

Madsen, F. W., Bank, T., Søndergaard, H. A. N., Mortensen, L. K., & Shaker, H. R. (2024). Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems: A Case Study in Denmark. In B. Nørregaard, L. C. Pereira da Silva, & Z. Ma (Eds.), Energy Informatics : Third Energy Informatics Academy Conference, EI.A 2023, Campinas, Brazil, December 6–8, 2023, Proceedings, Part II (Vol. 2, pp. 277–291). Springer. https://doi.org/10.1007/978-3-031-48652-4_18

Madsen, Frederik Wagner ; Bank, Theis ; Søndergaard, Henrik Alexander Nissen et al. / Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems : A Case Study in Denmark. Energy Informatics : Third Energy Informatics Academy Conference, EI.A 2023, Campinas, Brazil, December 6–8, 2023, Proceedings, Part II. editor / Bo Nørregaard ; Luiz Carlos Pereira da Silva ; Zheng Ma. Vol. 2 Springer, 2024. pp. 277–291 (Lecture Notes in Computer Science, Vol. 14468).

@inproceedings{28fd0fe0e2084dcda64885926a558787,

title = "Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems: A Case Study in Denmark",

abstract = "Faults in district heating systems (DHS) cause suboptimal operating conditions, which increase energy losses. As DHSs are critical infrastructure for many households in Denmark, these faults should be detected and corrected quickly. A novel model-based fault detection and diagnosis framework has been applied to detect and prioritise faults. The framework uses a bound for normal operation based on the residuals between historical sensor data and simulated properties in a digital twin of the DHS. The faults detected are prioritised based on the fault probability calculated using the Chernoff bound method. A case study on a Danish DHS has proven that the framework can produce a prioritised list of faults that maintenance crews can use to target faults with the highest probability. Furthermore, the digital twin allowed for fault location investigation, which could correlate different faults in the DHS. The framework has the potential for real-time fault detection and diagnosis. However, more precise digital twins need to be developed.",

keywords = "fault detection and diagnosis, district heating systems, digital twin, Chernoff bound",

author = "Madsen, {Frederik Wagner} and Theis Bank and S{\o}ndergaard, {Henrik Alexander Nissen} and Mortensen, {Lasse Kappel} and Shaker, {Hamid Reza}",

year = "2024",

doi = "10.1007/978-3-031-48652-4_18",

language = "English",

isbn = "978-3-031-48651-7",

volume = "2",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "277–291",

editor = "Bo N{\o}rregaard and {Pereira da Silva}, {Luiz Carlos} and Zheng Ma",

booktitle = "Energy Informatics",

address = "Germany",

note = "Energy Informatics.Academy Conference 2023, EI.A 2023 ; Conference date: 06-12-2023 Through 08-12-2023",

url = "https://www.energyinformatics.academy/eia-2023-conference",

}

Madsen, FW, Bank, T, Søndergaard, HAN, Mortensen, LK & Shaker, HR 2024, Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems: A Case Study in Denmark. in B Nørregaard, LC Pereira da Silva & Z Ma (eds), Energy Informatics : Third Energy Informatics Academy Conference, EI.A 2023, Campinas, Brazil, December 6–8, 2023, Proceedings, Part II. vol. 2, Springer, Lecture Notes in Computer Science, vol. 14468, pp. 277–291, Energy Informatics.Academy Conference 2023, São Paulo , Brazil, 06/12/2023. https://doi.org/10.1007/978-3-031-48652-4_18

Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems: A Case Study in Denmark. / Madsen, Frederik Wagner; Bank, Theis; Søndergaard, Henrik Alexander Nissen et al.
Energy Informatics : Third Energy Informatics Academy Conference, EI.A 2023, Campinas, Brazil, December 6–8, 2023, Proceedings, Part II. ed. / Bo Nørregaard; Luiz Carlos Pereira da Silva; Zheng Ma. Vol. 2 Springer, 2024. p. 277–291 (Lecture Notes in Computer Science, Vol. 14468).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems

T2 - Energy Informatics.Academy Conference 2023

AU - Madsen, Frederik Wagner

AU - Bank, Theis

AU - Søndergaard, Henrik Alexander Nissen

AU - Mortensen, Lasse Kappel

AU - Shaker, Hamid Reza

N1 - Conference code: 3

PY - 2024

Y1 - 2024

N2 - Faults in district heating systems (DHS) cause suboptimal operating conditions, which increase energy losses. As DHSs are critical infrastructure for many households in Denmark, these faults should be detected and corrected quickly. A novel model-based fault detection and diagnosis framework has been applied to detect and prioritise faults. The framework uses a bound for normal operation based on the residuals between historical sensor data and simulated properties in a digital twin of the DHS. The faults detected are prioritised based on the fault probability calculated using the Chernoff bound method. A case study on a Danish DHS has proven that the framework can produce a prioritised list of faults that maintenance crews can use to target faults with the highest probability. Furthermore, the digital twin allowed for fault location investigation, which could correlate different faults in the DHS. The framework has the potential for real-time fault detection and diagnosis. However, more precise digital twins need to be developed.

AB - Faults in district heating systems (DHS) cause suboptimal operating conditions, which increase energy losses. As DHSs are critical infrastructure for many households in Denmark, these faults should be detected and corrected quickly. A novel model-based fault detection and diagnosis framework has been applied to detect and prioritise faults. The framework uses a bound for normal operation based on the residuals between historical sensor data and simulated properties in a digital twin of the DHS. The faults detected are prioritised based on the fault probability calculated using the Chernoff bound method. A case study on a Danish DHS has proven that the framework can produce a prioritised list of faults that maintenance crews can use to target faults with the highest probability. Furthermore, the digital twin allowed for fault location investigation, which could correlate different faults in the DHS. The framework has the potential for real-time fault detection and diagnosis. However, more precise digital twins need to be developed.

KW - fault detection and diagnosis

KW - district heating systems

KW - digital twin

KW - Chernoff bound

U2 - 10.1007/978-3-031-48652-4_18

DO - 10.1007/978-3-031-48652-4_18

M3 - Article in proceedings

SN - 978-3-031-48651-7

VL - 2

T3 - Lecture Notes in Computer Science

SP - 277

EP - 291

BT - Energy Informatics

A2 - Nørregaard, Bo

A2 - Pereira da Silva, Luiz Carlos

A2 - Ma, Zheng

PB - Springer

Y2 - 6 December 2023 through 8 December 2023

ER -

Madsen FW, Bank T, Søndergaard HAN, Mortensen LK, Shaker HR. Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems: A Case Study in Denmark. In Nørregaard B, Pereira da Silva LC, Ma Z, editors, Energy Informatics : Third Energy Informatics Academy Conference, EI.A 2023, Campinas, Brazil, December 6–8, 2023, Proceedings, Part II. Vol. 2. Springer. 2024. p. 277–291. (Lecture Notes in Computer Science, Vol. 14468). doi: 10.1007/978-3-031-48652-4_18

Digital Twin-Based Fault Detection and Prioritisation in District Heating Systems: A Case Study in Denmark

Abstract

Conference

Keywords

Documents & Links

Fingerprint

Cite this