The buildings sector accounts for a significant portion of global energy consumption and greenhouse gas emissions, highlighting the need for sustainable and energy-efficient building design and construction techniques. This thesis proposes a parametric workflow that aims to mitigate the environmental impact of buildings by intervening in the early stages of design, demonstrating its functionality through the refurbishment of an existing building. The proposed approach aims to enhance the existing state-of-the-art practices by providing a real-time overview of a building's performance during the early design phases. The workflow incorporates an innovative method that leverages the capabilities of an artificial neural network, which is trained to predict three key performance indicators of the building: Spatial Daylight Autonomy (sDA), Annual Sunlight Exposure (ASE), and on-site energy production from solar panels. The predictions are reported in a user friendly interface that enables designers and clients to explore various design solutions while effectively showcasing a comprehensive view of the building's performance and providing insightful information on how the building responds to geometric changes. The thesis also investigates the use of Intelligent Details to streamline the design phase and enhance productivity during the production stage. The proposed solution is compared with traditional design and workflows, which often encounter challenges in promoting efficiency in the construction process, leading to fragmented projects. The primary objective of this thesis is to present an efficient workflow that would enable the attainment of the net-zero emissions target by 2050. The employed methodology appears to pave the way for enhancing the extant design practices and furnishing valuable insights into how sustainable design is no longer a mere possibility but an ineluctable need.
Il settore delle costruzioni rappresenta una parte significativa del consumo globale di energia e delle emissioni di gas serra, evidenziando la necessità di progettare e costruire edifici sostenibili ed energeticamente efficienti. Questa tesi propone un flusso di lavoro parametrico che mira a mitigare l'impatto ambientale degli edifici intervenendo nelle prime fasi del progetto, dimostrandone la funzionalità attraverso la ristrutturazione di un edificio esistente. L'approccio proposto mira a migliorare le pratiche esistenti fornendo una panoramica in tempo reale delle prestazioni dell'edificio durante le prime fasi del progetto. Il flusso di lavoro incorpora un metodo innovativo che sfrutta le capacità di una rete neurale artificiale, addestrata per prevedere tre indicatori chiave per le prestazioni dell'edificio: l'autonomia della luce diurna (sDA), l'esposizione annuale alla luce solare (ASE) e la produzione di energia in loco tramite pannelli solari. Le previsioni sono riportate in un'interfaccia utente intuitiva che consente ai progettisti e ai clienti di esplorare diverse soluzioni di progetto, mostrando efficacemente le prestazioni dell'edificio e fornendo informazioni utili su come l'edificio risponde ai cambiamenti geometrici. La tesi esplora anche l'uso di dettagli intelligenti per ottimizzare la fase di progettazione e aumentare la produttività durante la fase di produzione. La soluzione proposta viene confrontata con i flussi di lavoro di progettazione tradizionali, che spesso incontrano difficoltà nel promuovere l'efficienza nel processo di costruzione, portando a progetti frammentati. L'obiettivo principale di questa tesi è presentare un flusso di lavoro efficiente che consenta di raggiungere l'obiettivo di emissioni zero nette entro il 2050. La metodologia impiegata sembra aprire la strada per migliorare le pratiche di progettazione esistenti e fornire informazioni preziose su come il design sostenibile non sia più solo una possibilità ma una necessità ineluttabile.
From prediction to construction : a parametric workflow for volumetric facade panels. Integrating artificial neural networks and smart details to optimize buillding solutions
Legesse, Redait Tsegaye;LOSHKOV, KONSTANTIN;Azzone, Giandomenico
2022/2023
Abstract
The buildings sector accounts for a significant portion of global energy consumption and greenhouse gas emissions, highlighting the need for sustainable and energy-efficient building design and construction techniques. This thesis proposes a parametric workflow that aims to mitigate the environmental impact of buildings by intervening in the early stages of design, demonstrating its functionality through the refurbishment of an existing building. The proposed approach aims to enhance the existing state-of-the-art practices by providing a real-time overview of a building's performance during the early design phases. The workflow incorporates an innovative method that leverages the capabilities of an artificial neural network, which is trained to predict three key performance indicators of the building: Spatial Daylight Autonomy (sDA), Annual Sunlight Exposure (ASE), and on-site energy production from solar panels. The predictions are reported in a user friendly interface that enables designers and clients to explore various design solutions while effectively showcasing a comprehensive view of the building's performance and providing insightful information on how the building responds to geometric changes. The thesis also investigates the use of Intelligent Details to streamline the design phase and enhance productivity during the production stage. The proposed solution is compared with traditional design and workflows, which often encounter challenges in promoting efficiency in the construction process, leading to fragmented projects. The primary objective of this thesis is to present an efficient workflow that would enable the attainment of the net-zero emissions target by 2050. The employed methodology appears to pave the way for enhancing the extant design practices and furnishing valuable insights into how sustainable design is no longer a mere possibility but an ineluctable need.| File | Dimensione | Formato | |
|---|---|---|---|
|
2023_05_Azzone_Legesse_Loshkov.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: A4 Thesis book
Dimensione
27.23 MB
Formato
Adobe PDF
|
27.23 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/10589/210697