Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model

Kathrin Kefer; Samuel Haijes; Michael Mörth; Richard Heimrath; Thomas Mach; Valentin Kaisermayer; Christopher Zemann; Daniel Muschick; Bogdan Burlacu; Stephan Winkler; Michael Affenzeller

doi:10.46354/i3m.2023.sesde.004

Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model

Kathrin Kefer, Samuel Haijes, Michael Mörth, Richard Heimrath, Thomas Mach, Valentin Kaisermayer, Christopher Zemann, Daniel Muschick, Bogdan Burlacu, Stephan Winkler, Michael Affenzeller

Publikation: Beitrag in Buch/Bericht/Konferenzband › Beitrag in einem Konferenzband › Begutachtung

Abstract

Intelligent energy management systems can play a vital role in supporting themuch needed energy transition. However, in order to train machine learningmodels for this task, often very complex and detailed simulationmodels are needed. This can make the overall training process very slow or even impossible, which is why using resource efficient surrogates of the original simulationmodel during the training can be a possible solution. This work therefore focuses on the training of surrogates of a very detailed building simulation model using three different algorithms (k-Nearest Neighbour, RandomForest and Genetic Algorithm) and evaluates and compares them for their prediction capabilities, learned behaviours as well as execution time. Results show that the RandomForest algorithm achieves the best overall performance for 28 of the 35 surrogates, can learn the expected behavior and improves the execution speed by a factor of up to 664 compared to the original IDA ICE simulationmodel.

Originalsprache	englisch
Titel	11th International Workshop on Simulation for Energy, Sustainable Development and Environment, SESDE 2023
Redakteure/-innen	Agostino G. Bruzzone, Janos Sebestyen Janosy, Letizia Nicoletti, Gregory Zacharewicz
Seitenumfang	10
Band	23
ISBN (elektronisch)	978-888574198-0
DOIs	https://doi.org/10.46354/i3m.2023.sesde.004
Publikationsstatus	Veröffentlicht - Sept. 2023
Veranstaltung	11th International Workshop on Simulation for Energy, Sustainable Development and Environment: SESDE 2023 - Athens, Griechenland Dauer: 18 Sept. 2023 → 20 Sept. 2023

Konferenz

Konferenz	11th International Workshop on Simulation for Energy, Sustainable Development and Environment
Land/Gebiet	Griechenland
Ort	Athens
Zeitraum	18/09/23 → 20/09/23

Schlagwörter

Gebäudesimulation
Machine Learning
Energiemanagement

ASJC Scopus subject areas

Steuerungs- und Systemtechnik
Erneuerbare Energien, Nachhaltigkeit und Umwelt
Modellierung und Simulation

Fields of Expertise

Sustainable Systems

Treatment code (Nähere Zuordnung)

Experimental

UN SDGs

Dieser Output leistet einen Beitrag zu folgendem(n) Ziel(en) für nachhaltige Entwicklung

Zugriff auf Dokument

10.46354/i3m.2023.sesde.004

Andere Dateien und Links

Verknüpfung zur Publikation in Scopus

Dieses zitieren

Kefer, K., Haijes, S., Mörth, M., Heimrath, R., Mach, T., Kaisermayer, V., Zemann, C., Muschick, D., Burlacu, B., Winkler, S., & Affenzeller, M. (2023). Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model. in A. G. Bruzzone, J. S. Janosy, L. Nicoletti, & G. Zacharewicz (Hrsg.), 11th International Workshop on Simulation for Energy, Sustainable Development and Environment, SESDE 2023 (Band 23) https://doi.org/10.46354/i3m.2023.sesde.004

Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model. / Kefer, Kathrin; Haijes, Samuel; Mörth, Michael et al.
11th International Workshop on Simulation for Energy, Sustainable Development and Environment, SESDE 2023. Hrsg. / Agostino G. Bruzzone; Janos Sebestyen Janosy; Letizia Nicoletti; Gregory Zacharewicz. Band 23 2023.

Publikation: Beitrag in Buch/Bericht/Konferenzband › Beitrag in einem Konferenzband › Begutachtung

Kefer, K, Haijes, S, Mörth, M, Heimrath, R , Mach, T, Kaisermayer, V, Zemann, C, Muschick, D, Burlacu, B, Winkler, S & Affenzeller, M 2023, Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model. in AG Bruzzone, JS Janosy, L Nicoletti & G Zacharewicz (Hrsg.), 11th International Workshop on Simulation for Energy, Sustainable Development and Environment, SESDE 2023. Bd. 23, 11th International Workshop on Simulation for Energy, Sustainable Development and Environment, Athens, Griechenland, 18/09/23. https://doi.org/10.46354/i3m.2023.sesde.004

Kefer K, Haijes S, Mörth M, Heimrath R , Mach T, Kaisermayer V et al. Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model. in Bruzzone AG, Janosy JS, Nicoletti L, Zacharewicz G, Hrsg., 11th International Workshop on Simulation for Energy, Sustainable Development and Environment, SESDE 2023. Band 23. 2023 doi: 10.46354/i3m.2023.sesde.004

Kefer, Kathrin ; Haijes, Samuel ; Mörth, Michael et al. / Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model. 11th International Workshop on Simulation for Energy, Sustainable Development and Environment, SESDE 2023. Hrsg. / Agostino G. Bruzzone ; Janos Sebestyen Janosy ; Letizia Nicoletti ; Gregory Zacharewicz. Band 23 2023.

@inproceedings{4f4ddcf5647f478e93b373afe0d02ab8,

title = "Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model",

abstract = "Intelligent energy management systems can play a vital role in supporting themuch needed energy transition. However, in order to train machine learningmodels for this task, often very complex and detailed simulationmodels are needed. This can make the overall training process very slow or even impossible, which is why using resource efficient surrogates of the original simulationmodel during the training can be a possible solution. This work therefore focuses on the training of surrogates of a very detailed building simulation model using three different algorithms (k-Nearest Neighbour, RandomForest and Genetic Algorithm) and evaluates and compares them for their prediction capabilities, learned behaviours as well as execution time. Results show that the RandomForest algorithm achieves the best overall performance for 28 of the 35 surrogates, can learn the expected behavior and improves the execution speed by a factor of up to 664 compared to the original IDA ICE simulationmodel.",

keywords = "Geb{\"a}udesimulation, Machine Learning, Energiemanagement, Building Simulation Model Surrogates, Energy Management System, Heuristic Optimization, Machine Learning",

author = "Kathrin Kefer and Samuel Haijes and Michael M{\"o}rth and Richard Heimrath and Thomas Mach and Valentin Kaisermayer and Christopher Zemann and Daniel Muschick and Bogdan Burlacu and Stephan Winkler and Michael Affenzeller",

year = "2023",

month = sep,

doi = "10.46354/i3m.2023.sesde.004",

language = "English",

volume = "23",

editor = "Bruzzone, {Agostino G.} and Janosy, {Janos Sebestyen} and Letizia Nicoletti and Gregory Zacharewicz",

booktitle = "11th International Workshop on Simulation for Energy, Sustainable Development and Environment, SESDE 2023",

note = "11th International Workshop on Simulation for Energy, Sustainable Development and Environment : SESDE 2023 ; Conference date: 18-09-2023 Through 20-09-2023",

}

TY - GEN

T1 - Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model

AU - Kefer, Kathrin

AU - Haijes, Samuel

AU - Mörth, Michael

AU - Heimrath, Richard

AU - Mach, Thomas

AU - Kaisermayer, Valentin

AU - Zemann, Christopher

AU - Muschick, Daniel

AU - Burlacu, Bogdan

AU - Winkler, Stephan

AU - Affenzeller, Michael

PY - 2023/9

Y1 - 2023/9

N2 - Intelligent energy management systems can play a vital role in supporting themuch needed energy transition. However, in order to train machine learningmodels for this task, often very complex and detailed simulationmodels are needed. This can make the overall training process very slow or even impossible, which is why using resource efficient surrogates of the original simulationmodel during the training can be a possible solution. This work therefore focuses on the training of surrogates of a very detailed building simulation model using three different algorithms (k-Nearest Neighbour, RandomForest and Genetic Algorithm) and evaluates and compares them for their prediction capabilities, learned behaviours as well as execution time. Results show that the RandomForest algorithm achieves the best overall performance for 28 of the 35 surrogates, can learn the expected behavior and improves the execution speed by a factor of up to 664 compared to the original IDA ICE simulationmodel.

AB - Intelligent energy management systems can play a vital role in supporting themuch needed energy transition. However, in order to train machine learningmodels for this task, often very complex and detailed simulationmodels are needed. This can make the overall training process very slow or even impossible, which is why using resource efficient surrogates of the original simulationmodel during the training can be a possible solution. This work therefore focuses on the training of surrogates of a very detailed building simulation model using three different algorithms (k-Nearest Neighbour, RandomForest and Genetic Algorithm) and evaluates and compares them for their prediction capabilities, learned behaviours as well as execution time. Results show that the RandomForest algorithm achieves the best overall performance for 28 of the 35 surrogates, can learn the expected behavior and improves the execution speed by a factor of up to 664 compared to the original IDA ICE simulationmodel.

KW - Gebäudesimulation

KW - Machine Learning

KW - Energiemanagement

KW - Building Simulation Model Surrogates

KW - Energy Management System

KW - Heuristic Optimization

KW - Machine Learning

UR - http://www.scopus.com/inward/record.url?scp=85179134118&partnerID=8YFLogxK

U2 - 10.46354/i3m.2023.sesde.004

DO - 10.46354/i3m.2023.sesde.004

M3 - Conference paper

VL - 23

BT - 11th International Workshop on Simulation for Energy, Sustainable Development and Environment, SESDE 2023

A2 - Bruzzone, Agostino G.

A2 - Janosy, Janos Sebestyen

A2 - Nicoletti, Letizia

A2 - Zacharewicz, Gregory

T2 - 11th International Workshop on Simulation for Energy, Sustainable Development and Environment

Y2 - 18 September 2023 through 20 September 2023

ER -

Evaluating Machine Learning and Heuristic Optimization Based Surrogates as a Replacement for a Complex Building Simulation Model

Abstract

Konferenz

Schlagwörter

ASJC Scopus subject areas

Fields of Expertise

Treatment code (Nähere Zuordnung)

UN SDGs

Zugriff auf Dokument

Andere Dateien und Links

Fingerprint

Dieses zitieren