Evaluation of Energy and Lighting in Schools Considering Solar Protections DOI

Cristián Muñoz-Viveros,

Alexis Pérez-Fargallo, Carlos Rubio-Bellido

et al.

Published: Jan. 1, 2023

The implementation of the solar protections laid out in regulations and certification systems decreases cooling demands, although it has an uncontrolled even unforeseen effect on daylighting. Achieving a balance between both requirements is challenge for facade design, energy behavior, lighting performance since restricting radiation contributions daylight's contribution. This research aims to define methodology that allows, early design stage, evaluation protection solutions consider optimal daylight using annual dynamic indicators while maintaining adequate energy-saving levels. For this purpose, consumption have been considered as primary indicators, namely Spatial autonomy (sDA), Annual access (ASE), useful illuminance (UDI) with variations Modified Solar Factor (MSF). Thermal light assessments were made modeling two types protection. case study school classroom located city Talca, central Chile, considering window-to-wall ratios (WWR) 40%, 50%, 60%. premise's thermal behavior obtained sDA UDI results allowed making approximation solutions, however, ASE values all cases, classified unsuitable use. analysis suggests better limit are by organizing WWR instead MSF each solution identifies. compared options at reaching recommended levels savings 70% or more, It concluded achieve minimum acceptable daylighting levels, consumption, necessary relevant indicators.

Language: Английский

Guest editorial: Embracing the future construction project lifecycle: education and training for construction 4.0 DOI
Ali Rashidi, Mina Najafi, Mehrdad Arashpour

et al.

Smart and Sustainable Built Environment, Journal Year: 2024, Volume and Issue: 13(4), P. 757 - 762

Published: July 4, 2024

Language: Английский

Citations

2
Application of multimodal learning in daylight provision and view quality assessment of residential building layouts DOI

Charlotte Jeline Kat,

Fatemeh Mostafavi, Eleonora Brembilla

et al.

International Journal of Architectural Computing, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 24, 2024

Optimizing the layout of residential buildings based on daylight performance and view quality is crucial to visual comfort well-being building occupants. Machine Learning (ML) methods offer valuable support for performance-based decision-making process at early-stage design. In this study, a novel workflow introduced integrate ML models into architectural design process. With designer’s input floor designs, presented multimodal model predicts provision quality, which are then translated practical representations by post-processing step. This approach allows designs be evaluated model, leading enhanced decisions while preserving autonomy. Results best-performing implementing ResNet50 fully connected network, led Mean Square Error (MSE) 0.0440 0.0478, an R2 score 0.7411 0.7815 metrics, respectively. The results predictive further interpreted according different apartment categories various resolutions. These indicate that method could viable predicting in early tools, providing designers with faster feedback supports informed during iterations. Ultimately, challenges study improvements discussed.

Language: Английский

Citations

2
Architectural alchemy: Leveraging Artificial Intelligence for inspired design – a comprehensive study of creativity, control, and collaboration DOI Creative Commons
Lenka Petráková, Vladimír Šimkovič

Architecture Papers of the Faculty of Architecture and Design STU, Journal Year: 2023, Volume and Issue: 28(4), P. 3 - 14

Published: Dec. 1, 2023

Abstract The research paper contends that Artificial Intelligence (AI) serves as a collaborative partner in architectural design, rather than merely utility tool. To substantiate this argument, three-phase, nine-test investigation evaluating the strengths and limitations of two prominent AI platforms: Midjourney Stable Diffusion was undertaken. These platforms synergize human creativity capabilities through features like text prompts image references, thereby fostering innovative avenues architecture. Our analysis indicates is proficient generating initial design concepts, largely thanks to its extensive data libraries, but deficient refinement user control. Conversely, empowers designers with greater control via ControlNet sacrifices visual clarity due smaller generative models. Both share common flaw: an overemphasis on aesthetics shape at expense functional understanding. Building upon these empirical observations, outlines strategies for reasonably leverage optimising workflows. It confirms key hypotheses concerning interplay creativity, control, collaboration, emphasising both architects systems benefit from iterative feedback continuous adaptation. In summary, study posits not just adjunct technology transformative force capacity fundamentally alter processes, paving way new paradigm where expertise machine converge enriched outcomes.

Language: Английский

Citations

4
Ai-Acd: A Proposed Workflow of Ai-Powered Architectural Computational Design Process DOI

Basma Nashaat Ali,

Mostafa M. Elzeni

Published: Jan. 1, 2024

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

Language: Английский

Citations

1
Exploring the Potential of Generative Adversarial Networks in Enhancing Urban Renewal Efficiency DOI Open Access

Yunfei Lin,

Mingzhu Song

Sustainability, Journal Year: 2024, Volume and Issue: 16(13), P. 5768 - 5768

Published: July 6, 2024

As Chinese cities transition into a stage of stock development, the revitalization industrial areas becomes increasingly crucial, serving as pivotal factor in urban renewal. The renovation old factory buildings is full swing, and architects often rely on matured experience to produce several profile schemes for selection during process. However, when dealing with large number factories, this task can consume significant amount manpower. In era maturing machine learning, study, set against backdrop an district, explores potential application deep learning technology improving efficiency renovation. We establish generation model based generative adversarial networks (GANs), generating design features building profiles. To ensure balance between feasibility creativity generated designs, study employs various transformation techniques each original image dataset construction, creating mappings images schemes. Additionally, data augmentation are applied expand dataset, trained models validated analyzed test set. This demonstrates GANs design, providing designers richer reference solutions.

Language: Английский

Citations

1
Generative AI models for different steps in architectural design: A literature review DOI Creative Commons
Chengyuan Li, Tianyu Zhang,

Xusheng Du

et al.

Frontiers of Architectural Research, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 1, 2024

Language: Английский

Citations

1
The Development of an Experimental Framework to Explore the Generative Design Preference of a Machine Learning-Assisted Residential Site Plan Layout DOI Creative Commons

Pei Sun,

Fengying Yan,

Qiwei He

et al.

Land, Journal Year: 2023, Volume and Issue: 12(9), P. 1776 - 1776

Published: Sept. 13, 2023

Generative design based on machine learning has become an important area of application for artificial intelligence. Regarding the generative process residential site plan layouts (hereafter referred to as “RSPLs”), lack experimental demonstration begs question: what are preferences learning? In this case, all elements target object need be extracted much possible conduct studies produce scientific results. Based this, Pix2pix model was used test case Chinese areas in study. An framework “extract-translate-machine-learning-evaluate” is proposed, combining different and manual computations, well quantitative qualitative evaluation techniques, jointly determine which their characteristic representations field RSPL. The results show that can assist optimizing two particular RSPL conform layout plans: plaza paving landscaped green space. addition, other major elements, public facilities spatial structures, were also found exhibit more significant preferences, with largest percentage increase number changes required after learning. Finally, established study compensates consideration a simultaneously utilize same methodological framework. This planners developing solutions better meet expectations residents clarify potential advantageous directions learning-assisted

Language: Английский

Citations

3
Evaluation of energy and lighting in schools considering solar protections DOI Creative Commons

Cristián Muñoz-Viveros,

Carlos Rubio-Bellido, Alexis Pérez-Fargallo

et al.

Energy Reports, Journal Year: 2024, Volume and Issue: 12, P. 1157 - 1177

Published: July 18, 2024

Language: Английский

Citations

0
Evaluation and Optimization of Traditional Mountain Village Spatial Environment Performance Using Genetic and XGBoost Algorithms in the Early Design Stage—A Case Study in the Cold Regions of China DOI Creative Commons
Zhixin Xu, Xiaoming Li, Bo Sun

et al.

Buildings, Journal Year: 2024, Volume and Issue: 14(9), P. 2796 - 2796

Published: Sept. 5, 2024

As urbanization advances, rural construction and resource development in China encounter significant challenges, leading to the widespread adoption of standardized planning design methods manage increasing population pressure. These uniform approaches often prioritize economic benefits over climate adaptability energy efficiency. This paper addresses this issue by focusing on traditional mountain villages northern regions, particularly examining wind thermal environments courtyards street networks. study integrates consumption comfort performance analysis early process, utilizing Genetic XGBoost algorithms enhance began selecting a benchmark model based simulations courtyard PET (Physiological Equivalent Temperature) MRT (mean radiant temperature). It then employed Wallacei_X plugin, which uses NSGA-II algorithm for multi-objective genetic optimization (MOGO) optimize five objectives. The resulting solutions were trained Scikit-learn machine learning platform. After comparing models like RandomForest XGBoost, highest-performing was selected further training. Validation shows that achieves an average accuracy 80% predicting performance. In project’s validation phase, overall network framework block first adjusted prediction related strategies. optimized prototype integrated into scheme according functional requirements. repeated adjustments, village conducted. calculations indicate improves 36% compared with original baseline. conclusion, aimed integrate assessment decision-making process optimizing environments, offering new sustainable development.

Language: Английский

Citations

0
Enhancing visual comfort prediction in space layouts using deep learning models DOI
Amirali Ghourkhanehchi Zirak, Afshin Saeedian, Zahra Sadat Zomorodian

et al.

Intelligent Buildings International, Journal Year: 2023, Volume and Issue: 15(6), P. 266 - 284

Published: Nov. 2, 2023

In designing performance-based facilities, such as educational buildings, assessing visual comfort is crucial. The computational cost and resource-intensive modeling necessitate an efficient analytical tool for diverse layout assessment. the current study, predictive models were developed using deep learning machine to predict in typical elementary schools Tehran. Firstly, layouts modeled, considering influential parameters. Subsequently, simulation-based datasets analyzed labeled. VGG16 VGG19 Architectures from convolutional neural networks (CNNs), along with pix2pix model generative adversarial forecasting respectively, numerical pictorial indices regarding each metric. performs approximately SSIM of 0.9 sDA, ASE, UDIexceed3000 0.86 DAp. extracted features by CNNs harnessed training models. Eventually, Bayesian Ridge algorithm had a promising performance which exhibited acceptable R2 values metrics range 0.90–0.96. Toward depth analysis, parameter sensitivity Shapley additive explanations method was evaluated XGBoost Additionally, Spearman Correlation underscores substantial impact factors like WWR aspect ratio on metrics.

Language: Английский

Citations

1