Modeling of Industrial Hot-Water System within SSAB Borlänge
(2025) MVKM05 20251Department of Energy Sciences
- Abstract
- In this thesis, the modeling and simulation of SSAB’s hot water system in Borlänge was studied based on system descriptions and existing data. In simple terms, this system represents the internal district heating network of SSAB, which supplies hot water to the buildings of the facilities, both for heating and warm water consumption. While some amount of the produced heat is consumed by processes and the municipality’s district heating system, Borlänge Energi, the main focus lies on the heat consumption of the substations (UCs). The main characteristic of this study was the individual representation of each substation, as a contrast to some literature where the heat consumers were treated as one large unit. However, limiting factors such... (More)
- In this thesis, the modeling and simulation of SSAB’s hot water system in Borlänge was studied based on system descriptions and existing data. In simple terms, this system represents the internal district heating network of SSAB, which supplies hot water to the buildings of the facilities, both for heating and warm water consumption. While some amount of the produced heat is consumed by processes and the municipality’s district heating system, Borlänge Energi, the main focus lies on the heat consumption of the substations (UCs). The main characteristic of this study was the individual representation of each substation, as a contrast to some literature where the heat consumers were treated as one large unit. However, limiting factors such as missing temperatures of the building’s circulating streams, as well as assumptions of uniform heat exchanger characteristics where data was absent, resulted in challenges for simulating certain parameters accurately. For instance, a coefficient of determination of 0.27 indicated low model performance in predicting the variability of the primary return temperature to Energi Central 1, despite capturing fairly accurately the average values (CV-RMSE=8.3 %). These findings underscore the critical influence of data gaps on model fidelity for such complex systems and highlight that while the chosen software, Ebsilon®Professional, is proficient for component-based thermodynamic modeling, robust network performance prediction heavily relies on comprehensive operational data. (Less)
- Abstract (Swedish)
- I detta examensarbete studerades modellering och simulering av SSAB:s hetvattensystem i Borlänge baserat på systembeskrivningar och befintliga data. Enkelt uttryckt representerar detta system SSAB:s interna fjärrvärmenät, vilket försörjer anläggningens byggnader med hetvatten för både uppvärmning och tappvarmvatten. Även om en del av den producerade värmen förbrukas av processer och av kommunens fjärrvärmesystem, Borlänge Energi, ligger huvudfokus på värmeförbrukningen i undercentralerna (UC). Det huvudsakliga kännetecknet för denna studie var den individuella representationen av varje undercentral, i kontrast till viss litteratur där värmeförbrukarna behandlades som en enda stor enhet. Begränsande faktorer, såsom saknade temperaturer för... (More)
- I detta examensarbete studerades modellering och simulering av SSAB:s hetvattensystem i Borlänge baserat på systembeskrivningar och befintliga data. Enkelt uttryckt representerar detta system SSAB:s interna fjärrvärmenät, vilket försörjer anläggningens byggnader med hetvatten för både uppvärmning och tappvarmvatten. Även om en del av den producerade värmen förbrukas av processer och av kommunens fjärrvärmesystem, Borlänge Energi, ligger huvudfokus på värmeförbrukningen i undercentralerna (UC). Det huvudsakliga kännetecknet för denna studie var den individuella representationen av varje undercentral, i kontrast till viss litteratur där värmeförbrukarna behandlades som en enda stor enhet. Begränsande faktorer, såsom saknade temperaturer för byggnadernas cirkulerande flöden samt antaganden om enhetliga värmeväxlaregenskaper där data saknades, ledde dock till utmaningar med att simulera vissa parametrar noggrant. Till exempel indikerade en determinationskoefficient på 0,27 låg modellprestanda vid prediktering av variabiliteten i den primära returtemperaturen till Energi Central 1, trots att medelvärdena fångades relativt noggrant (CV-RMSE=8,3 %). Dessa resultat understryker det kritiska inflytandet av dataluckor på modelltillförlitligheten för sådana komplexa system och belyser att medan den valda programvaran, Ebsilon®Professional, är kompetent för komponentbaserad termodynamisk modellering, är robust prediktering av nätverksprestanda starkt beroende av omfattande driftdata. (Less)
- Abstract (Greek, Modern (1453-))
- Σε αυτήν τη διπλωματική εργασία, μελετήθηκε η μοντελοποίηση και προσομοίωση του συστήματος ζεστού νερού της SSAB στο Borlänge, βάσει περιγραφών του συστήματος και υφιστάμενων δεδομένων. Με απλούς όρους, αυτό το σύστημα αποτελεί το εσωτερικό δίκτυο τηλεθέρμανσης της SSAB, το οποίο παρέχει ζεστό νερό στα κτίρια των εγκαταστάσεων, τόσο για θέρμανση όσο και για κατανάλωση ζεστού νερού χρήσης. Ενώ ένα μέρος της παραγόμενης θερμότητας καταναλώνεται από παραγωγικές διαδικασίες και το σύστημα τηλεθέρμανσης του δήμου, Borlänge Energi, το επίκεντρο έγκειται στην κατανάλωση θερμότητας των υποσταθμών (UCs). Το κύριο χαρακτηριστικό αυτής της μελέτης ήταν η μεμονωμένη αναπαράσταση κάθε υποσταθμού, σε αντίθεση με μέρος της βιβλιογραφίας όπου οι... (More)
- Σε αυτήν τη διπλωματική εργασία, μελετήθηκε η μοντελοποίηση και προσομοίωση του συστήματος ζεστού νερού της SSAB στο Borlänge, βάσει περιγραφών του συστήματος και υφιστάμενων δεδομένων. Με απλούς όρους, αυτό το σύστημα αποτελεί το εσωτερικό δίκτυο τηλεθέρμανσης της SSAB, το οποίο παρέχει ζεστό νερό στα κτίρια των εγκαταστάσεων, τόσο για θέρμανση όσο και για κατανάλωση ζεστού νερού χρήσης. Ενώ ένα μέρος της παραγόμενης θερμότητας καταναλώνεται από παραγωγικές διαδικασίες και το σύστημα τηλεθέρμανσης του δήμου, Borlänge Energi, το επίκεντρο έγκειται στην κατανάλωση θερμότητας των υποσταθμών (UCs). Το κύριο χαρακτηριστικό αυτής της μελέτης ήταν η μεμονωμένη αναπαράσταση κάθε υποσταθμού, σε αντίθεση με μέρος της βιβλιογραφίας όπου οι καταναλωτές θερμότητας αντιμετωπίζονταν ως μία ενιαία μεγάλη μονάδα. Ωστόσο, περιοριστικοί παράγοντες όπως η έλλειψη δεδομένων για τις θερμοκρασίες των ρευμάτων νερού των κτιρίων, καθώς και παραδοχές για ομοιόμορφα χαρακτηριστικά των εναλλακτών θερμότητας όπου τα δεδομένα απουσίαζαν, οδήγησαν σε προκλήσεις για την ακριβή προσομοίωση ορισμένων παραμέτρων. Για παράδειγμα, ο συντελεστής προσδιορισμού (R2) με τιμή 0,27 υπέδειξε χαμηλή απόδοση του μοντέλου στην πρόβλεψη της μεταβλητότητας της θερμοκρασίας επιστροφής του πρωτεύοντος δικτύου προς τον σταθμό Energi Central 1 (EC1), παρά την αρκετά ακριβή αποτύπωση των μέσων τιμών (CV-RMSE=8,3 %). Αυτά τα ευρήματα υπογραμμίζουν την κρίσιμη επίδραση των κενών στα δεδομένα στην εγκυρότητα του μοντέλου για τέτοια πολύπλοκα συστήματα και τονίζουν ότι, ενώ το επιλεγμένο λογισμικό, Ebsilon®Professional, είναι ικανό για θερμοδυναμική μοντελοποίηση βασισμένη στα επιμέρους στοιχεία, η αξιόπιστη πρόβλεψη της απόδοσης του δικτύου εξαρτάται σε μεγάλο βαθμό από την ύπαρξη ολοκληρωμένων λειτουργικών δεδομένων. (Less)
- Popular Abstract
- How do you fine-tune a massive industrial heating system without shutting it down? You build a "digital twin". This project took on the challenge of creating a complete computer model of the vast hot-water network at the SSAB steel plant in Borlänge. The goal was to map its energy flows and find hidden ways to make it even better. The system itself is a clever design of recycling, capturing waste heat from steelmaking to warm the plant's buildings and even help heat local and municipal homes. But its sheer size and complexity make it tough to understand completely.
Let's face it, steelmaking uses a tremendous amount of energy and has a big carbon footprint, responsible for about 7-9% of global CO₂ emissions. That’s why companies like... (More) - How do you fine-tune a massive industrial heating system without shutting it down? You build a "digital twin". This project took on the challenge of creating a complete computer model of the vast hot-water network at the SSAB steel plant in Borlänge. The goal was to map its energy flows and find hidden ways to make it even better. The system itself is a clever design of recycling, capturing waste heat from steelmaking to warm the plant's buildings and even help heat local and municipal homes. But its sheer size and complexity make it tough to understand completely.
Let's face it, steelmaking uses a tremendous amount of energy and has a big carbon footprint, responsible for about 7-9% of global CO₂ emissions. That’s why companies like SSAB are aiming to be fossil-free by 2045. A huge part of that journey is getting smart about waste heat. The Borlänge system is a perfect example. It works like a district heating network, grabbing heat from industrial processes and boilers and using it to heat buildings across the site. It's so effective, it even sells the extra heat to the city of Borlänge.
To build this digital twin, the project used software called Ebsilon®Professional, which lets you construct a virtual system piece by piece. Every boiler, pipe, and heat exchanger was put into the model as a digital block. The trickiest part was figuring out how much heat each of the 43 buildings used, because most of them didn't have detailed sensors. To get around this, the model had to make educated guesses using old data from 2020, linking heat use to the outdoor temperature.
So, how did the digital twin do when tested against real-world data? It had pros and cons. The good news: the model was great at predicting the average temperature of the water returning to the main heating plant, with its predictions rated as "good". It also got the general seasonal trends right. The bad news: it struggled to predict the real-time ups and downs. While the average was on point, the model missed the sudden spikes and dips that happen in the real world. The reason came down to one thing: the lack of live data from the buildings forced the model to oversimplify how they use heat.
This project highlights a golden rule in the digital world: a model is only as good as its data. But this digital twin is far from pointless. It’s a vital first step, providing SSAB with a foundational tool to see how its massive heating system actually behaves. The path forward is clear: add more sensors to get a more detailed, live picture of what's happening. With better data, this digital twin could evolve into a powerful tool, opening the door for smart, possibly AI-driven controls that could help SSAB save energy and push forward on its journey to fossil-free steelmaking. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9199647
- author
- Kourtidis, Angelos LU
- supervisor
- organization
- course
- MVKM05 20251
- year
- 2025
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- Industrial hot-water system, district heating, system simulation, model validation, thermodynamic modeling
- report number
- ISRN LUTMDN/TMPH-25/5630-SE
- ISSN
- 0282-1990
- language
- English
- id
- 9199647
- date added to LUP
- 2025-06-16 14:11:14
- date last changed
- 2025-06-16 14:11:14
@misc{9199647,
abstract = {{In this thesis, the modeling and simulation of SSAB’s hot water system in Borlänge was studied based on system descriptions and existing data. In simple terms, this system represents the internal district heating network of SSAB, which supplies hot water to the buildings of the facilities, both for heating and warm water consumption. While some amount of the produced heat is consumed by processes and the municipality’s district heating system, Borlänge Energi, the main focus lies on the heat consumption of the substations (UCs). The main characteristic of this study was the individual representation of each substation, as a contrast to some literature where the heat consumers were treated as one large unit. However, limiting factors such as missing temperatures of the building’s circulating streams, as well as assumptions of uniform heat exchanger characteristics where data was absent, resulted in challenges for simulating certain parameters accurately. For instance, a coefficient of determination of 0.27 indicated low model performance in predicting the variability of the primary return temperature to Energi Central 1, despite capturing fairly accurately the average values (CV-RMSE=8.3 %). These findings underscore the critical influence of data gaps on model fidelity for such complex systems and highlight that while the chosen software, Ebsilon®Professional, is proficient for component-based thermodynamic modeling, robust network performance prediction heavily relies on comprehensive operational data.}},
author = {{Kourtidis, Angelos}},
issn = {{0282-1990}},
language = {{eng}},
note = {{Student Paper}},
title = {{Modeling of Industrial Hot-Water System within SSAB Borlänge}},
year = {{2025}},
}