In EDS ansehen

Drones and community-powered rapid neighborhood energy modeling: Demonstrated in a real-world case study.

Gespeichert in:
Bibliographische Detailangaben
Titel: Drones and community-powered rapid neighborhood energy modeling: Demonstrated in a real-world case study.
Autoren: Mohammed, Noushad Ahamed Chittoor1,2 (AUTHOR), Adesanya, Misbaudeen Aderemi1,2 (AUTHOR), Chowdhury, SoumyaDeep1,2 (AUTHOR), Debnath, Sudipta1,2 (AUTHOR), Halliday, Andrew3,4 (AUTHOR), Randhawa, Gurjit S.5 (AUTHOR), Farooque, Aitazaz A.2,6 (AUTHOR), Grewal, Kuljeet Singh1,2 (AUTHOR) kgrewal@upei.ca
Quelle: Sustainable Cities & Society. Jan2026, Vol. 136, pN.PAG-N.PAG. 1p.
Schlagwörter: Drone aircraft, Community involvement, Multisensor data fusion, Machine learning, Urban planning, Sustainability, Clean energy
Geografische Kategorien: Prince Edward Island
Abstract: • Introduced RENSA, a rapid energy modeling of neighborhoods through split automation. • RENSA enables scalable and high-precision neighborhood energy modeling. • Combines drone imagery, LiDAR, and ML with community-sourced data. • Community-engaged workflow enhances model realism and access to energy planning. • Provides a replicable framework for sustainable and net-zero urban future. Accurate evaluation of energy performance in neighborhood energy modeling (NEM) has long been challenged by the lack of representative ground-truth data and high-resolution 3D geometry. Traditional approaches often rely on simplified shoebox archetypes, semantic 3D city models, and standardized assumptions for occupancy, internal loads, and envelope characteristics, thereby limiting model fidelity. To overcome these constraints, this study presents a holistic approach to NEM by introducing rapid energy modeling of neighborhoods through split automation (RENSA) – a novel, semi-automated workflow that integrates community engagement, multi-source data fusion, and machine learning to generate various level of detail (LoD) building models from LoD0–LoD3 and enable high-precision, context-sensitive NEM simulations. RENSA is applied to 297 structures in Georgetown, Prince Edward Island (PE), Canada, producing LoD3 models for the entire community and conducting detailed energy simulations for 71 residential buildings with available utility data. A structured, JavaScript object notation (JSON)-based data pipeline automates simulation inputs gathered through community engagement. Buildings are classified into five energy system categories, and model outputs are validated against real utility records. Geometric validation shows mean absolute percentage errors of 4.27 % for footprint area (LoD0), 3.55 % for peak height (LoD1), 7.48 % for bottom chord height (LoD2), 6.80 % for volume (LoD2), and 12.88 % for fenestration areas (LoD3). Further, integrating community-led data into the calibration workflow allowed 65 % of the simulated buildings to meet the ASHRAE normalized mean bias error (NMBE) requirement of ±5 %. The RENSA generalized framework demonstrates a replicable, scalable, and community-driven approach to NEM, enabling user-defined LoD generation and effectively bridging the gap between theory and real-world application in support of net-zero energy transitions. [ABSTRACT FROM AUTHOR]
Datenbank: Supplemental Index
Beschreibung
Abstract:• Introduced RENSA, a rapid energy modeling of neighborhoods through split automation. • RENSA enables scalable and high-precision neighborhood energy modeling. • Combines drone imagery, LiDAR, and ML with community-sourced data. • Community-engaged workflow enhances model realism and access to energy planning. • Provides a replicable framework for sustainable and net-zero urban future. Accurate evaluation of energy performance in neighborhood energy modeling (NEM) has long been challenged by the lack of representative ground-truth data and high-resolution 3D geometry. Traditional approaches often rely on simplified shoebox archetypes, semantic 3D city models, and standardized assumptions for occupancy, internal loads, and envelope characteristics, thereby limiting model fidelity. To overcome these constraints, this study presents a holistic approach to NEM by introducing rapid energy modeling of neighborhoods through split automation (RENSA) – a novel, semi-automated workflow that integrates community engagement, multi-source data fusion, and machine learning to generate various level of detail (LoD) building models from LoD0–LoD3 and enable high-precision, context-sensitive NEM simulations. RENSA is applied to 297 structures in Georgetown, Prince Edward Island (PE), Canada, producing LoD3 models for the entire community and conducting detailed energy simulations for 71 residential buildings with available utility data. A structured, JavaScript object notation (JSON)-based data pipeline automates simulation inputs gathered through community engagement. Buildings are classified into five energy system categories, and model outputs are validated against real utility records. Geometric validation shows mean absolute percentage errors of 4.27 % for footprint area (LoD0), 3.55 % for peak height (LoD1), 7.48 % for bottom chord height (LoD2), 6.80 % for volume (LoD2), and 12.88 % for fenestration areas (LoD3). Further, integrating community-led data into the calibration workflow allowed 65 % of the simulated buildings to meet the ASHRAE normalized mean bias error (NMBE) requirement of ±5 %. The RENSA generalized framework demonstrates a replicable, scalable, and community-driven approach to NEM, enabling user-defined LoD generation and effectively bridging the gap between theory and real-world application in support of net-zero energy transitions. [ABSTRACT FROM AUTHOR]
ISSN:22106707
DOI:10.1016/j.scs.2025.107089