Artificial Neural Networks to assess energy and environmental performance of buildings: An Italian case study

Valerio Lo Brano, Giuseppina Ciulla, Antonino D'Amico, Marzia Traverso, D'Amico, Ciulla, Palumbo, Valerio Lo Brano

Risultato della ricerca: Article

1 Citazione (Scopus)

Abstract

Approximately 40% of the European energy consumption and a large proportion of environmental impacts are related to the building sector. However, the selection of adequate and correct designs can provide considerable energy savings and reduce environmental impacts. To achieve this objective, a simultaneous energy and environmental assessment of a building's life cycle is necessary. To date, the resolution of this complex problem is entrusted to numerous software and calculation algorithms that are often complex to use. They involve long diagnosis phases and are characterised by the lack of a common language. Despite the efforts by the scientific community in the building sector, there is no simple and reliable tool that simultaneously solves the energy and environmental balance of buildings. In this work, the authors address this challenge by proposing the application of an Artificial Neural Network. Due to the high reliability of learning algorithms in the resolution of complex and non-linear problems, it was possible to simultaneously solve two different but strongly dependent aspects after a deep training phase. In previous researches, the authors applied several topologies of neural networks,which were trained on a large and representative database and developed for the Italian building stock.The database, characterised by several building models simulated in different climatic conditions, collects 29 inputs (13 energy data and 16 environmental data) and provides 7 outputs, 1 for heating energy demand and 6 of the most used indicators in life cycle assessment of buildings. A statistical analysis of the results confirmed that the proposed method is appropriate to achieve the goal of the study. The best artificial neural network for each output presented low Root Mean Square Error, Mean Absolute Error lower than 5%, and determination coefficient close to 1. The excellent results confirmed that this methodology can be extended in any context and to any condition (other countries and building stocks).Furthermore, the implementation of this solution algorithm in a software program can enable the development of a suitable decision support tool, which is simple, reliable, and easy to use even for a non-expert user. The possibility to use an instrument to predict a building's performance in its design and planning phase, represent an important result to support decision-making processes toward more sustainable choices.
Lingua originaleEnglish
Numero di pagine21
RivistaJOURNAL OF CLEANER PRODUCTION
Volume239
Stato di pubblicazionePublished - 2019

Fingerprint

artificial neural network
Neural networks
environmental impact
life cycle
Environmental impact
energy
software
Life cycle
environmental assessment
topology
energy balance
statistical analysis
learning
Mean square error
Learning algorithms
heating
Statistical methods
Energy conservation
Energy utilization
methodology

All Science Journal Classification (ASJC) codes

  • Renewable Energy, Sustainability and the Environment
  • Environmental Science(all)
  • Strategy and Management
  • Industrial and Manufacturing Engineering

Cita questo

@article{7f6119eec042413aaa89a38b20d6be2c,
title = "Artificial Neural Networks to assess energy and environmental performance of buildings: An Italian case study",
abstract = "Approximately 40{\%} of the European energy consumption and a large proportion of environmental impacts are related to the building sector. However, the selection of adequate and correct designs can provide considerable energy savings and reduce environmental impacts. To achieve this objective, a simultaneous energy and environmental assessment of a building's life cycle is necessary. To date, the resolution of this complex problem is entrusted to numerous software and calculation algorithms that are often complex to use. They involve long diagnosis phases and are characterised by the lack of a common language. Despite the efforts by the scientific community in the building sector, there is no simple and reliable tool that simultaneously solves the energy and environmental balance of buildings. In this work, the authors address this challenge by proposing the application of an Artificial Neural Network. Due to the high reliability of learning algorithms in the resolution of complex and non-linear problems, it was possible to simultaneously solve two different but strongly dependent aspects after a deep training phase. In previous researches, the authors applied several topologies of neural networks,which were trained on a large and representative database and developed for the Italian building stock.The database, characterised by several building models simulated in different climatic conditions, collects 29 inputs (13 energy data and 16 environmental data) and provides 7 outputs, 1 for heating energy demand and 6 of the most used indicators in life cycle assessment of buildings. A statistical analysis of the results confirmed that the proposed method is appropriate to achieve the goal of the study. The best artificial neural network for each output presented low Root Mean Square Error, Mean Absolute Error lower than 5{\%}, and determination coefficient close to 1. The excellent results confirmed that this methodology can be extended in any context and to any condition (other countries and building stocks).Furthermore, the implementation of this solution algorithm in a software program can enable the development of a suitable decision support tool, which is simple, reliable, and easy to use even for a non-expert user. The possibility to use an instrument to predict a building's performance in its design and planning phase, represent an important result to support decision-making processes toward more sustainable choices.",
author = "{Lo Brano}, Valerio and Giuseppina Ciulla and Antonino D'Amico and Marzia Traverso and D'Amico and Ciulla and Palumbo and {Lo Brano}, Valerio",
year = "2019",
language = "English",
volume = "239",
journal = "JOURNAL OF CLEANER PRODUCTION",
issn = "0959-6526",

}

TY - JOUR

T1 - Artificial Neural Networks to assess energy and environmental performance of buildings: An Italian case study

AU - Lo Brano, Valerio

AU - Ciulla, Giuseppina

AU - D'Amico, Antonino

AU - Traverso, Marzia

AU - D'Amico, null

AU - Ciulla, null

AU - Palumbo, null

AU - Lo Brano, Valerio

PY - 2019

Y1 - 2019

N2 - Approximately 40% of the European energy consumption and a large proportion of environmental impacts are related to the building sector. However, the selection of adequate and correct designs can provide considerable energy savings and reduce environmental impacts. To achieve this objective, a simultaneous energy and environmental assessment of a building's life cycle is necessary. To date, the resolution of this complex problem is entrusted to numerous software and calculation algorithms that are often complex to use. They involve long diagnosis phases and are characterised by the lack of a common language. Despite the efforts by the scientific community in the building sector, there is no simple and reliable tool that simultaneously solves the energy and environmental balance of buildings. In this work, the authors address this challenge by proposing the application of an Artificial Neural Network. Due to the high reliability of learning algorithms in the resolution of complex and non-linear problems, it was possible to simultaneously solve two different but strongly dependent aspects after a deep training phase. In previous researches, the authors applied several topologies of neural networks,which were trained on a large and representative database and developed for the Italian building stock.The database, characterised by several building models simulated in different climatic conditions, collects 29 inputs (13 energy data and 16 environmental data) and provides 7 outputs, 1 for heating energy demand and 6 of the most used indicators in life cycle assessment of buildings. A statistical analysis of the results confirmed that the proposed method is appropriate to achieve the goal of the study. The best artificial neural network for each output presented low Root Mean Square Error, Mean Absolute Error lower than 5%, and determination coefficient close to 1. The excellent results confirmed that this methodology can be extended in any context and to any condition (other countries and building stocks).Furthermore, the implementation of this solution algorithm in a software program can enable the development of a suitable decision support tool, which is simple, reliable, and easy to use even for a non-expert user. The possibility to use an instrument to predict a building's performance in its design and planning phase, represent an important result to support decision-making processes toward more sustainable choices.

AB - Approximately 40% of the European energy consumption and a large proportion of environmental impacts are related to the building sector. However, the selection of adequate and correct designs can provide considerable energy savings and reduce environmental impacts. To achieve this objective, a simultaneous energy and environmental assessment of a building's life cycle is necessary. To date, the resolution of this complex problem is entrusted to numerous software and calculation algorithms that are often complex to use. They involve long diagnosis phases and are characterised by the lack of a common language. Despite the efforts by the scientific community in the building sector, there is no simple and reliable tool that simultaneously solves the energy and environmental balance of buildings. In this work, the authors address this challenge by proposing the application of an Artificial Neural Network. Due to the high reliability of learning algorithms in the resolution of complex and non-linear problems, it was possible to simultaneously solve two different but strongly dependent aspects after a deep training phase. In previous researches, the authors applied several topologies of neural networks,which were trained on a large and representative database and developed for the Italian building stock.The database, characterised by several building models simulated in different climatic conditions, collects 29 inputs (13 energy data and 16 environmental data) and provides 7 outputs, 1 for heating energy demand and 6 of the most used indicators in life cycle assessment of buildings. A statistical analysis of the results confirmed that the proposed method is appropriate to achieve the goal of the study. The best artificial neural network for each output presented low Root Mean Square Error, Mean Absolute Error lower than 5%, and determination coefficient close to 1. The excellent results confirmed that this methodology can be extended in any context and to any condition (other countries and building stocks).Furthermore, the implementation of this solution algorithm in a software program can enable the development of a suitable decision support tool, which is simple, reliable, and easy to use even for a non-expert user. The possibility to use an instrument to predict a building's performance in its design and planning phase, represent an important result to support decision-making processes toward more sustainable choices.

UR - http://hdl.handle.net/10447/369125

M3 - Article

VL - 239

JO - JOURNAL OF CLEANER PRODUCTION

JF - JOURNAL OF CLEANER PRODUCTION

SN - 0959-6526

ER -