Cities and their neighbourhoods are responsible for a significant portion of global energy consumption and greenhouse gas emissions, posing urgent environmental challenges. Current initiatives to reduce the carbon footprint of cities are ham-pered by numerous factors, including the complexity of energy assessments, the lack of effective tools in the early stages of design, and the difficulty of considering urban systems as interconnected sets of buildings rather than isolated structures. In this context, the role of designers is crucial to promote technical and social responses to climate challenges. It is essential to equip professionals with accessible and intuitive methods to enable them to assess the environmental impact of buil-dings without burdening costs and without limiting design creativity, ensuring sufficiently thorough assessments without oversimplification. In view of these considerations, the thesis accompanies the development of a framework that facilitates the integration of sustainability into urban projects from the earliest stages of development, through the use of data-driven technologies and artificial intelligence machine learning algorithms. The proposal of a decision-making tool that can be easily integrated into existing workflows and is capable of producing an accurate picture of consumption represents a methodological and practical advance for the sector. The developed workflow is distinguished by its ability to integrate tools that can quantitatively simulate the energy and environmental performance of entire neighbourhoods, rather than individual buildings. This model aims to provide a practical solution that allows planners to optimise energy efficiency, reducing the time and resources required for analysis, while ensuring a high level of accuracy and detail. Within the thesis, calculations and elaborations were developed to support the development of this design framework, which was then verified through a specific case study: the design of a zero-emission district in Iceland, in Reykjavik, by se-lecting a lot from the international competition ‘Reinventing Cities’ promoted by C40. The project constitutes a reference model for the application of large-scale decarbonisation strategies and was an ideal testing ground for the framework, con-sidering Iceland’s climate challenges. The use of Machine Learning models, which have been tested and validated against Business-As-Usual tools, made it possible to quickly simulate different configurations on various project scales, facilitating the identification of the most effective solutions for reducing emissions. By introducing this quantitative approach for decarbonisation planning, it provides a concrete solution to the lack of inte-grated tools and scalable models for assessing and improving energy performance early in the design phase, when there is a limited amount of information. The simplicity of the framework in terms of data requirements and its speed of execution represent a significant step forward compared to traditional methods, which often require detailed data and long proces-sing times, thus limiting the possibility of exploring sustainable solutions. The importance of this work lies, therefore, in its ability to provide practitioners with an innovative and versatile design support tool. Thanks to a data-driven approach based on artificial intelligence algorithms, the proposed framework en-courages sustainable design decisions, responding to the growing need for practical and accessible tools to reduce the environmental impact of cities. The developed case study demonstrates the applicability of the framework and its benefits in the real-world context, concretely paving the way for more responsible urban design aligned with global climate goals.
Le città e i loro quartieri sono responsabili di una parte rilevante del consumo di energia e delle emissioni globali di gas serra, ponendo sfide urgenti sul piano ambientale. Le iniziative attuali per la riduzione dell’impronta di carbonio delle città sono ostacolate da numerosi fattori, tra cui la complessità delle valutazioni energetiche, la mancanza di strumenti efficaci nelle fasi iniziali di progettazione e la difficoltà di considerare i sistemi urbani come insiemi di edifici interconnessi, anziché come strutture isolate. In questo contesto, il ruolo dei progettisti è determinante per promuovere risposte tecniche e sociali alle sfide climatiche. È essenziale dotare i professionisti di metodi accessibili e intuitivi per consentire loro di valutare l’impatto ambientale degli edifici senza gravare sui costi e senza limitare la creatività progettuale, garantendo valutazioni sufficientemente approfon-dite senza eccessive semplificazioni. Alla luce di queste considerazioni, la tesi accompagna lo sviluppo di un framework che agevoli l’integrazione della soste-nibilità nei progetti urbani fin dalle prime fasi di sviluppo, tramite l’utilizzo di tecnologie data-driven e algoritmi Machine Learning di intelligenza artificiale. La proposta di uno strumento decisionale facilmente integrabile nei flussi di lavoro esistenti e capace di produrre un qua-dro accurato dei consumi rappresenta un avanzamento metodologico e pratico per il settore. Il flusso di lavoro elaborato si distingue per la capacità di integrare strumenti in grado di simulare, in modo quantitativo, le prestazioni energetiche e ambientali di interi quartieri, piuttosto che di singoli edifici. Questo modello punta a fornire una soluzione pratica che permetta ai progettisti di ottimizzare l’efficienza energetica, riducendo il tempo e le risorse necessarie per l’analisi, e garan-tendo contemporaneamente un alto livello di precisione e dettaglio. Nel quadro della tesi sono stati sviluppati calcoli ed elaborazioni a supporto dello sviluppo di questo framework proget-tuale, che è stato poi verificato attraverso un caso studio specifico: la progettazione di un quartiere a zero emissioni in Islanda, a Reykjavik, selezionando un lotto dal concorso internazionale “Reinventing Cities” promosso da C40. Il progetto costituisce un modello di riferimento per l’applicazione di strategie di decarbonizzazione su larga scala e ha rappresentato un banco di prova ideale per il framework, considerando le sfide climatiche islandesi. L’utilizzo di modelli Machine Learning, testati e validati rispetto a strumenti Business-As-Usual, ha permesso di simulare in modo rapido diverse configurazioni su varie scale di progetto, facilitando l’identificazione delle soluzioni più efficaci per la riduzione delle emissioni. Introducendo questo approccio quantitativo per la pianificazione della decarbonizzazione, si fornisce una soluzione con-creta alla mancanza di strumenti integrati e di modelli scalabili per valutare e migliorare le prestazioni energetiche fin dalle prime fasi di progettazione, quando si ha una quantità ridotta di informazioni. La semplicità del framework in termini di re-quisiti di dati e la sua velocità di esecuzione rappresentano un passo avanti significativo rispetto ai metodi tradizionali, che spesso richiedono dati dettagliati e lunghe tempistiche di elaborazione, limitando così la possibilità di esplorare soluzioni sostenibili. L’importanza di questo lavoro risiede, quindi, nella capacità di fornire ai professionisti del settore uno strumento innovativo e versatile di supporto alla progettazione. Grazie a un approccio data-driven, basato su algoritmi di intelligenza artificiale, il framework proposto incoraggia decisioni progettuali sostenibili, rispondendo alla crescente esigenza di strumenti, pratici e accessibili, per ridurre l’impatto ambientale delle città. Il caso studio sviluppato dimostra l’applicabilità del framework e i suoi vantaggi nel contesto reale, aprendo concretamente la strada a una progettazione urbana più responsabile e allineata con gli obiettivi climatici globali.
AI-enabled design framework for a carbon-neutral neighbourhood: a case study in Reykjavik
Viganò, Riccardo;Aliprandi, Alessandro
2023/2024
Abstract
Cities and their neighbourhoods are responsible for a significant portion of global energy consumption and greenhouse gas emissions, posing urgent environmental challenges. Current initiatives to reduce the carbon footprint of cities are ham-pered by numerous factors, including the complexity of energy assessments, the lack of effective tools in the early stages of design, and the difficulty of considering urban systems as interconnected sets of buildings rather than isolated structures. In this context, the role of designers is crucial to promote technical and social responses to climate challenges. It is essential to equip professionals with accessible and intuitive methods to enable them to assess the environmental impact of buil-dings without burdening costs and without limiting design creativity, ensuring sufficiently thorough assessments without oversimplification. In view of these considerations, the thesis accompanies the development of a framework that facilitates the integration of sustainability into urban projects from the earliest stages of development, through the use of data-driven technologies and artificial intelligence machine learning algorithms. The proposal of a decision-making tool that can be easily integrated into existing workflows and is capable of producing an accurate picture of consumption represents a methodological and practical advance for the sector. The developed workflow is distinguished by its ability to integrate tools that can quantitatively simulate the energy and environmental performance of entire neighbourhoods, rather than individual buildings. This model aims to provide a practical solution that allows planners to optimise energy efficiency, reducing the time and resources required for analysis, while ensuring a high level of accuracy and detail. Within the thesis, calculations and elaborations were developed to support the development of this design framework, which was then verified through a specific case study: the design of a zero-emission district in Iceland, in Reykjavik, by se-lecting a lot from the international competition ‘Reinventing Cities’ promoted by C40. The project constitutes a reference model for the application of large-scale decarbonisation strategies and was an ideal testing ground for the framework, con-sidering Iceland’s climate challenges. The use of Machine Learning models, which have been tested and validated against Business-As-Usual tools, made it possible to quickly simulate different configurations on various project scales, facilitating the identification of the most effective solutions for reducing emissions. By introducing this quantitative approach for decarbonisation planning, it provides a concrete solution to the lack of inte-grated tools and scalable models for assessing and improving energy performance early in the design phase, when there is a limited amount of information. The simplicity of the framework in terms of data requirements and its speed of execution represent a significant step forward compared to traditional methods, which often require detailed data and long proces-sing times, thus limiting the possibility of exploring sustainable solutions. The importance of this work lies, therefore, in its ability to provide practitioners with an innovative and versatile design support tool. Thanks to a data-driven approach based on artificial intelligence algorithms, the proposed framework en-courages sustainable design decisions, responding to the growing need for practical and accessible tools to reduce the environmental impact of cities. The developed case study demonstrates the applicability of the framework and its benefits in the real-world context, concretely paving the way for more responsible urban design aligned with global climate goals.File | Dimensione | Formato | |
---|---|---|---|
2024_12_Aliprandi_Viganò.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
101.5 MB
Formato
Adobe PDF
|
101.5 MB | Adobe PDF | Visualizza/Apri |
2024_12_Aliprandi_Viganò_Elaborati.pdf
accessibile in internet per tutti
Descrizione: Elaborati grafici
Dimensione
69.89 MB
Formato
Adobe PDF
|
69.89 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/231252