A new zone temperature predictive modeling for energy saving in buildings

Hao Huang, Lei Chen, Morteza Mohammadzaheri, Eric Hu

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

Currently in most buildings, the heating, ventilation and air conditioning (HVAC) systems are controlled by the present temperature in the building. If the predictions for future temperature in the building or a zone were available, the building management system (BMS) could use both present and future temperatures to control HVAC systems, the energy consumed by HAVC systems could then be minimised. Therefore, a lot of research effort has been devoted to develop accurate temperature prediction models using various approaches, e.g. traditional thermodynamic, artificial neural networks (ANN), generic algorithms (GA) and fuzzy logic approaches. When the historical data of the building is available, the ANN approach is thought to be the most cost-effective method. Most of previous studies of ANN modelling of building temperature, have either focused on single-zone examination or assumed that zones' temperatures were the same throughout the building. In this study, a more realistic multi-zone scenario in a large building is proposed in the developing of the ANN temperature predictive model. The coupled effects between zones caused by the temperature difference are considered in the model. The results of a case study show that the new ANN model that considers the temperatures of the neighbouring zones, achieves more accurate results. The proposed modelling methodology can be extended to include other inputs, besides neighboring zones' temperatures, usage pattern of the building, so that the better intelligent control strategies can be developed for energy saving purposes, based on the more accurate predicted temperatures form the new model.

Original languageEnglish
Pages (from-to)142-151
Number of pages10
JournalProcedia Engineering
Volume49
DOIs
Publication statusPublished - 2012

Fingerprint

Energy conservation
Temperature
Neural networks
Air conditioning
Ventilation
Heating
Intelligent control
Fuzzy logic
Thermodynamics

Keywords

  • Artificial neural networks (ANN)
  • HVAC
  • Multi-zone

ASJC Scopus subject areas

  • Engineering(all)

Cite this

A new zone temperature predictive modeling for energy saving in buildings. / Huang, Hao; Chen, Lei; Mohammadzaheri, Morteza; Hu, Eric.

In: Procedia Engineering, Vol. 49, 2012, p. 142-151.

Research output: Contribution to journalArticle

@article{47cbcfb7570a41f29e65a781fbc42702,
title = "A new zone temperature predictive modeling for energy saving in buildings",
abstract = "Currently in most buildings, the heating, ventilation and air conditioning (HVAC) systems are controlled by the present temperature in the building. If the predictions for future temperature in the building or a zone were available, the building management system (BMS) could use both present and future temperatures to control HVAC systems, the energy consumed by HAVC systems could then be minimised. Therefore, a lot of research effort has been devoted to develop accurate temperature prediction models using various approaches, e.g. traditional thermodynamic, artificial neural networks (ANN), generic algorithms (GA) and fuzzy logic approaches. When the historical data of the building is available, the ANN approach is thought to be the most cost-effective method. Most of previous studies of ANN modelling of building temperature, have either focused on single-zone examination or assumed that zones' temperatures were the same throughout the building. In this study, a more realistic multi-zone scenario in a large building is proposed in the developing of the ANN temperature predictive model. The coupled effects between zones caused by the temperature difference are considered in the model. The results of a case study show that the new ANN model that considers the temperatures of the neighbouring zones, achieves more accurate results. The proposed modelling methodology can be extended to include other inputs, besides neighboring zones' temperatures, usage pattern of the building, so that the better intelligent control strategies can be developed for energy saving purposes, based on the more accurate predicted temperatures form the new model.",
keywords = "Artificial neural networks (ANN), HVAC, Multi-zone",
author = "Hao Huang and Lei Chen and Morteza Mohammadzaheri and Eric Hu",
year = "2012",
doi = "10.1016/j.proeng.2012.10.122",
language = "English",
volume = "49",
pages = "142--151",
journal = "Procedia Engineering",
issn = "1877-7058",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - A new zone temperature predictive modeling for energy saving in buildings

AU - Huang, Hao

AU - Chen, Lei

AU - Mohammadzaheri, Morteza

AU - Hu, Eric

PY - 2012

Y1 - 2012

N2 - Currently in most buildings, the heating, ventilation and air conditioning (HVAC) systems are controlled by the present temperature in the building. If the predictions for future temperature in the building or a zone were available, the building management system (BMS) could use both present and future temperatures to control HVAC systems, the energy consumed by HAVC systems could then be minimised. Therefore, a lot of research effort has been devoted to develop accurate temperature prediction models using various approaches, e.g. traditional thermodynamic, artificial neural networks (ANN), generic algorithms (GA) and fuzzy logic approaches. When the historical data of the building is available, the ANN approach is thought to be the most cost-effective method. Most of previous studies of ANN modelling of building temperature, have either focused on single-zone examination or assumed that zones' temperatures were the same throughout the building. In this study, a more realistic multi-zone scenario in a large building is proposed in the developing of the ANN temperature predictive model. The coupled effects between zones caused by the temperature difference are considered in the model. The results of a case study show that the new ANN model that considers the temperatures of the neighbouring zones, achieves more accurate results. The proposed modelling methodology can be extended to include other inputs, besides neighboring zones' temperatures, usage pattern of the building, so that the better intelligent control strategies can be developed for energy saving purposes, based on the more accurate predicted temperatures form the new model.

AB - Currently in most buildings, the heating, ventilation and air conditioning (HVAC) systems are controlled by the present temperature in the building. If the predictions for future temperature in the building or a zone were available, the building management system (BMS) could use both present and future temperatures to control HVAC systems, the energy consumed by HAVC systems could then be minimised. Therefore, a lot of research effort has been devoted to develop accurate temperature prediction models using various approaches, e.g. traditional thermodynamic, artificial neural networks (ANN), generic algorithms (GA) and fuzzy logic approaches. When the historical data of the building is available, the ANN approach is thought to be the most cost-effective method. Most of previous studies of ANN modelling of building temperature, have either focused on single-zone examination or assumed that zones' temperatures were the same throughout the building. In this study, a more realistic multi-zone scenario in a large building is proposed in the developing of the ANN temperature predictive model. The coupled effects between zones caused by the temperature difference are considered in the model. The results of a case study show that the new ANN model that considers the temperatures of the neighbouring zones, achieves more accurate results. The proposed modelling methodology can be extended to include other inputs, besides neighboring zones' temperatures, usage pattern of the building, so that the better intelligent control strategies can be developed for energy saving purposes, based on the more accurate predicted temperatures form the new model.

KW - Artificial neural networks (ANN)

KW - HVAC

KW - Multi-zone

UR - http://www.scopus.com/inward/record.url?scp=84882290433&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84882290433&partnerID=8YFLogxK

U2 - 10.1016/j.proeng.2012.10.122

DO - 10.1016/j.proeng.2012.10.122

M3 - Article

VL - 49

SP - 142

EP - 151

JO - Procedia Engineering

JF - Procedia Engineering

SN - 1877-7058

ER -