A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Futureproofing Building Retrofits: Developing a Surrogate Model for Climate Change Adaptation
https://orcid.org/http://orcid.org/0000-0003-4977-5125
https://orcid.org/http://orcid.org/0000-0001-7177-727X
An innovative and multi-faceted approach is necessary to address the complexities of climate change and its profound impact on the building sector. This approach must meld a nuanced understanding of prevailing climatic trends with robust methodologies aimed at futureproofing building retrofit decisions.
Given their extended lifespans of 50–60 years, buildings emerge as primary entities in this environmental challenge, necessitating the development of comprehensive strategies to adapt to these climatic shifts over prolonged durations. However, the existing building energy modelling and retrofit analysis paradigms are beset with computational contingencies, which are further compounded when projecting future scenarios.
This research proposes a surrogate model to evaluate building performance against future weather scenarios. The model’s foundation is a single-building surrogate constructed utilizing the XGBoost algorithm, selected for its exceptional generalizability. This model is then integrated with morphed weather files corresponding to projected future climates, necessitating selecting a representative environment RCP scenario.
Central to this methodology is the application of feature engineering processes designed to amalgamate disparate data structures within a unified model. This is achieved by applying dimensionality reduction techniques, pivotal in distilling and learning critical features from temporal weather data. Manual extraction methods are also proposed to complement this.
Employing an autoencoder model facilitates the transformation of these features into a concise latent space representation. The resultant enriched surrogate model can encompass multiyear weather data with its augmented structure. This pivotal enhancement enables instantaneous predictions of future building performance, effectively mitigating the computational burdens associated with long-term assessments. The model’s design is responsive to current climatic conditions and exhibits resilience and adaptability to anticipated future climatic variations.
Futureproofing Building Retrofits: Developing a Surrogate Model for Climate Change Adaptation
https://orcid.org/http://orcid.org/0000-0003-4977-5125
https://orcid.org/http://orcid.org/0000-0001-7177-727X
An innovative and multi-faceted approach is necessary to address the complexities of climate change and its profound impact on the building sector. This approach must meld a nuanced understanding of prevailing climatic trends with robust methodologies aimed at futureproofing building retrofit decisions.
Given their extended lifespans of 50–60 years, buildings emerge as primary entities in this environmental challenge, necessitating the development of comprehensive strategies to adapt to these climatic shifts over prolonged durations. However, the existing building energy modelling and retrofit analysis paradigms are beset with computational contingencies, which are further compounded when projecting future scenarios.
This research proposes a surrogate model to evaluate building performance against future weather scenarios. The model’s foundation is a single-building surrogate constructed utilizing the XGBoost algorithm, selected for its exceptional generalizability. This model is then integrated with morphed weather files corresponding to projected future climates, necessitating selecting a representative environment RCP scenario.
Central to this methodology is the application of feature engineering processes designed to amalgamate disparate data structures within a unified model. This is achieved by applying dimensionality reduction techniques, pivotal in distilling and learning critical features from temporal weather data. Manual extraction methods are also proposed to complement this.
Employing an autoencoder model facilitates the transformation of these features into a concise latent space representation. The resultant enriched surrogate model can encompass multiyear weather data with its augmented structure. This pivotal enhancement enables instantaneous predictions of future building performance, effectively mitigating the computational burdens associated with long-term assessments. The model’s design is responsive to current climatic conditions and exhibits resilience and adaptability to anticipated future climatic variations.
Futureproofing Building Retrofits: Developing a Surrogate Model for Climate Change Adaptation
https://orcid.org/http://orcid.org/0000-0003-4977-5125
https://orcid.org/http://orcid.org/0000-0001-7177-727X
An innovative and multi-faceted approach is necessary to address the complexities of climate change and its profound impact on the building sector. This approach must meld a nuanced understanding of prevailing climatic trends with robust methodologies aimed at futureproofing building retrofit decisions.
Given their extended lifespans of 50–60 years, buildings emerge as primary entities in this environmental challenge, necessitating the development of comprehensive strategies to adapt to these climatic shifts over prolonged durations. However, the existing building energy modelling and retrofit analysis paradigms are beset with computational contingencies, which are further compounded when projecting future scenarios.
This research proposes a surrogate model to evaluate building performance against future weather scenarios. The model’s foundation is a single-building surrogate constructed utilizing the XGBoost algorithm, selected for its exceptional generalizability. This model is then integrated with morphed weather files corresponding to projected future climates, necessitating selecting a representative environment RCP scenario.
Central to this methodology is the application of feature engineering processes designed to amalgamate disparate data structures within a unified model. This is achieved by applying dimensionality reduction techniques, pivotal in distilling and learning critical features from temporal weather data. Manual extraction methods are also proposed to complement this.
Employing an autoencoder model facilitates the transformation of these features into a concise latent space representation. The resultant enriched surrogate model can encompass multiyear weather data with its augmented structure. This pivotal enhancement enables instantaneous predictions of future building performance, effectively mitigating the computational burdens associated with long-term assessments. The model’s design is responsive to current climatic conditions and exhibits resilience and adaptability to anticipated future climatic variations.
Futureproofing Building Retrofits: Developing a Surrogate Model for Climate Change Adaptation
Lecture Notes in Civil Engineering
Berardi, Umberto (editor) / Saad, Mostafa (author) / Eicker, Ursula (author)
International Association of Building Physics ; 2024 ; Toronto, ON, Canada
2024-12-19
8 pages
Article/Chapter (Book)
Electronic Resource
English
| British Library Online Contents | 2009
Whole-Building Retrofits: A Gateway to Climate Stabilization
| British Library Online Contents | 2010
| Taylor & Francis Verlag | 2023