A significant number of school buildings in Italy require seismic and energy retrofits based on National laws, which contribute to the school environment's characteristics and health and safety in buildings. Moreover, government initiatives to promote ambitious national plans for the renovation and construction of new school buildings are gaining vast attention. For this purpose, the Ministry of Education, with the local authorities' collaboration, carries out a database to register national school buildings and their level of consistency and functionality, which is the fundamental knowledge tool for planning interventions in the sector. However, it does not provide a guideline to estimate future interventions' costs. This research aims to design a retrofitting cost estimation model for energy and seismic improvement and adaptation interventions using Artificial Neural Networks. It can serve as a beneficial tool for forecasting expenses based on the interrelated building features, which the public administration can use to optimize the management and planning of school buildings' funds. The proposed work focuses on a small sample of over 200 school buildings and their seismic and energy retrofitting costs. The ANN model uses the parameters of the case studies as the input to train the network to estimate the retrofitting cost of other projects based on the historical data. The parameters are categorized into three groups of features: i) building’s characteristics, e.g., construction year and the number of floors, ii) energy retrofit parameters, e.g., class heating energy consumption, and iii) seismic retrofit parameters, e.g., seismic zone and structural type. Therefore, the goal is to facilitate the financial feasibility assessments and optimize the available resources related to the planning of interventions. The proposed model will contribute significantly to school buildings' resilience as a single integrated space, which has the characteristics of habitability, flexibility, functionality, comfort, and well-being.
Application of Machine Learning to estimate Retrofitting Cost of School Buildings
Panarelli Gianmichele;
In corso di stampa
Abstract
A significant number of school buildings in Italy require seismic and energy retrofits based on National laws, which contribute to the school environment's characteristics and health and safety in buildings. Moreover, government initiatives to promote ambitious national plans for the renovation and construction of new school buildings are gaining vast attention. For this purpose, the Ministry of Education, with the local authorities' collaboration, carries out a database to register national school buildings and their level of consistency and functionality, which is the fundamental knowledge tool for planning interventions in the sector. However, it does not provide a guideline to estimate future interventions' costs. This research aims to design a retrofitting cost estimation model for energy and seismic improvement and adaptation interventions using Artificial Neural Networks. It can serve as a beneficial tool for forecasting expenses based on the interrelated building features, which the public administration can use to optimize the management and planning of school buildings' funds. The proposed work focuses on a small sample of over 200 school buildings and their seismic and energy retrofitting costs. The ANN model uses the parameters of the case studies as the input to train the network to estimate the retrofitting cost of other projects based on the historical data. The parameters are categorized into three groups of features: i) building’s characteristics, e.g., construction year and the number of floors, ii) energy retrofit parameters, e.g., class heating energy consumption, and iii) seismic retrofit parameters, e.g., seismic zone and structural type. Therefore, the goal is to facilitate the financial feasibility assessments and optimize the available resources related to the planning of interventions. The proposed model will contribute significantly to school buildings' resilience as a single integrated space, which has the characteristics of habitability, flexibility, functionality, comfort, and well-being.File | Dimensione | Formato | |
---|---|---|---|
RRAU_conference_paper_Application of Machine Learning to estimate Retrofitting Cost of School Buildings.pdf
accesso aperto
Dimensione
592.8 kB
Formato
Adobe PDF
|
592.8 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.