Hainan Yan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hainan Yan has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Shiliang Wang
School of Civil Engineering and Architecture, University of Jinan, Jinan 250022, China
2 shared publications Baihui Zhang
School of Architecture and Urban Planning, Nanjing University, Nanjing, China
1 shared publication- Bart Julien Dewancker
Department of Architecture, The University of Kitakyushu, Kitakyushu, Japan
1 shared publication - Boyang Bai1 shared publication
- Daiyuan Zhang
School of Civil Engineering and Transportation, Yangzhou University, Yangzhou 225009, China
1 shared publication - Deqing Bu
School of Architecture and Art, North China University of Technology, Beijing 100144, China
1 shared publication
Latest Publications
- Open Access
Multi-Objective Optimization of Energy Performance, Winter Cold Discomfort, and Daylighting in a Contemporary Yi Rural House Prototype: A Case Study in Xichang, China
Policy-led housing renewal and standardized rural-house atlases are reshaping contemporary Yi rural housing, yet trade-offs among energy performance, winter cold discomfort, and daylighting remain poorly quantified. This study aimed to quantify the effects of design parameters on energy use, winter cold discomfort, and daylight availability, and to examine the trade-offs among these objectives in a contemporary single-storey three-bay Yi rural house prototype in Xichang. The parametric prototype was informed by field surveys, policy documents, and a local rural-house atlas. Building performance simulations generated 1667 valid cases for PMV-based cold discomfort intensity (PMV-CDI), useful daylight illuminance (UDI300–2000), and simulation-based conditioned-zone energy use intensity (EUI). Six machine-learning algorithms were compared to develop surrogate models, with all selected models achieving independent-test CV(RMSE) values below 2%. SHAP analysis identified distinct contributions of envelope, window, shading, and spatial features across the three objectives. Surrogate-assisted NSGA-III generated 388 empirical nondominated solutions forming a continuous approximate Pareto front. The Pareto front revealed diminishing daylight gains accompanied by increasingly pronounced winter cold discomfort toward the high-daylight region. Hierarchical clustering translated the continuous Pareto front into three interpretable performance-priority scenarios, enabling comparison of the objective trade-offs and associated parameter combinations. The resulting framework provides a quantitative basis for comparing performance priorities in standardized rural-house design while retaining locally defined spatial and construction constraints.
Open AccessMulti-Objective Optimization of Energy Performance, Winter Cold Discomfort, and Daylighting in a Contemporary Yi Rural House Prototype: A Case Study in Xichang, China
Policy-led housing renewal and standardized rural-house atlases are reshaping contemporary Yi rural housing, yet trade-offs among energy performance, winter cold discomfort, and daylighting remain poorly quantified. This study aimed to quantify the effects of design parameters on energy use, winter cold discomfort, and daylight availability, and to examine the trade-offs among these objectives in a contemporary single-storey three-bay Yi rural house prototype in Xichang. The parametric prototype was informed by field surveys, policy documents, and a local rural-house atlas. Building performance simulations generated 1667 valid cases for PMV-based cold discomfort intensity (PMV-CDI), useful daylight illuminance (UDI300–2000), and simulation-based conditioned-zone energy use intensity (EUI). Six machine-learning algorithms were compared to develop surrogate models, with all selected models achieving independent-test CV(RMSE) values below 2%. SHAP analysis identified distinct contributions of envelope, window, shading, and spatial features across the three objectives. Surrogate-assisted NSGA-III generated 388 empirical nondominated solutions forming a continuous approximate Pareto front. The Pareto front revealed diminishing daylight gains accompanied by increasingly pronounced winter cold discomfort toward the high-daylight region. Hierarchical clustering translated the continuous Pareto front into three interpretable performance-priority scenarios, enabling comparison of the objective trade-offs and associated parameter combinations. The resulting framework provides a quantitative basis for comparing performance priorities in standardized rural-house design while retaining locally defined spatial and construction constraints.
- Open Access
A Multimodal Deep Learning Framework for Rapid Prediction of Operational Carbon Emissions in Early-Stage Residential Building Design
This study introduces a domain-specific multimodal deep learning framework, centered on a Vision Transformer (ViT), to accelerate the prediction of operational carbon emissions in residential buildings. Our approach uniquely fuses two data modalities: the geometric information captured in floorplan images and the quantitative data from vector-based building parameters. By training and testing on a comprehensive dataset of 17,000 residential samples derived from a large-scale open-source Chinese database, the proposed model demonstrates exceptional predictive capability. On the test set, it achieved a mean bias error of 1.75%, a mean absolute percentage error of 2.14%, and a coefficient of determination (R2) of 0.95. Further validation through comparative analysis shows that our framework significantly outperforms established deep learning architectures, including ResNet-18, Inception-V4, and VGG-19, in both accuracy and robustness. The developed tool provides architects with a reliable and rapid method for assessing the carbon footprint of design options, thereby offering crucial scientific support for sustainable building design.
Open AccessA Multimodal Deep Learning Framework for Rapid Prediction of Operational Carbon Emissions in Early-Stage Residential Building Design
This study introduces a domain-specific multimodal deep learning framework, centered on a Vision Transformer (ViT), to accelerate the prediction of operational carbon emissions in residential buildings. Our approach uniquely fuses two data modalities: the geometric information captured in floorplan images and the quantitative data from vector-based building parameters. By training and testing on a comprehensive dataset of 17,000 residential samples derived from a large-scale open-source Chinese database, the proposed model demonstrates exceptional predictive capability. On the test set, it achieved a mean bias error of 1.75%, a mean absolute percentage error of 2.14%, and a coefficient of determination (R2) of 0.95. Further validation through comparative analysis shows that our framework significantly outperforms established deep learning architectures, including ResNet-18, Inception-V4, and VGG-19, in both accuracy and robustness. The developed tool provides architects with a reliable and rapid method for assessing the carbon footprint of design options, thereby offering crucial scientific support for sustainable building design.
- 29 August 2025
- Published by Elsevier in Journal of Building Engineering
- Vol. 112
Multimodal generative adversarial networks for accelerated daylight prediction in residential buildings
- 29 August 2025
- Published by Elsevier in Journal of Building Engineering
- Vol. 112
Multimodal generative adversarial networks for accelerated daylight prediction in residential buildings
- Open Access
Multi-Objective Optimization of Outdoor Thermal Comfort and Sunlight Hours in Elderly Residential Areas: A Case Study of Beijing, China
To optimize the outdoor thermal comfort and sunlight hours of elderly residential areas in cold regions of China, we collected data on streets and building forms from 121 elderly residential sites in Beijing. Utilizing parametric modeling tools to generate ideal residential models, a multi-objective optimization algorithm was applied to identify 144 Pareto solutions. The optimal solutions were analyzed using K-means clustering and Pearson correlation analysis to examine how block form affects outdoor environmental performance. The universal thermal climate index (UTCI) in summer showed significant positive correlations (r > 0.72) with the distance between buildings (DB), building density (BD), shape coefficient (SC), and coefficient of variation for building height (CVH), and significant negative correlations (r < −0.82) with average building height (AH), floor area ratio (FAR), volume area ratio (VAR), mean building area (MA), average building volume (AV), and open space ratio (OSR). Winter UTCI was significantly positively correlated with AH, FAR, VAR, MA, and AV (r > 0.83) and significantly negatively correlated with DB, porosity (PO), SC, and CVH (r < −0.88). Sunlight hours were significantly positively correlated with DB, PO, OSR, and CVH (r > 0.84) and significantly negatively correlated with AH, BD, FAR, SC, VAR, MA, and AV (r > 0.88). Courtyard and point-building configurations performed the best across all optimization objectives. (The value of r, Pearson’s correlation coefficient, ranges from −1 to +1. r = +1: Perfect positive correlation, r = −1: Perfect negative correlation, r = 0: No linear correlation).
Open AccessMulti-Objective Optimization of Outdoor Thermal Comfort and Sunlight Hours in Elderly Residential Areas: A Case Study of Beijing, China
To optimize the outdoor thermal comfort and sunlight hours of elderly residential areas in cold regions of China, we collected data on streets and building forms from 121 elderly residential sites in Beijing. Utilizing parametric modeling tools to generate ideal residential models, a multi-objective optimization algorithm was applied to identify 144 Pareto solutions. The optimal solutions were analyzed using K-means clustering and Pearson correlation analysis to examine how block form affects outdoor environmental performance. The universal thermal climate index (UTCI) in summer showed significant positive correlations (r > 0.72) with the distance between buildings (DB), building density (BD), shape coefficient (SC), and coefficient of variation for building height (CVH), and significant negative correlations (r < −0.82) with average building height (AH), floor area ratio (FAR), volume area ratio (VAR), mean building area (MA), average building volume (AV), and open space ratio (OSR). Winter UTCI was significantly positively correlated with AH, FAR, VAR, MA, and AV (r > 0.83) and significantly negatively correlated with DB, porosity (PO), SC, and CVH (r < −0.88). Sunlight hours were significantly positively correlated with DB, PO, OSR, and CVH (r > 0.84) and significantly negatively correlated with AH, BD, FAR, SC, VAR, MA, and AV (r > 0.88). Courtyard and point-building configurations performed the best across all optimization objectives. (The value of r, Pearson’s correlation coefficient, ranges from −1 to +1. r = +1: Perfect positive correlation, r = −1: Perfect negative correlation, r = 0: No linear correlation).
- 10 September 2024
- Published by Springer Nature in Building Simulation
- Vol. 17 (11) , 2125-2143
Developing an integrated prediction model for daylighting, thermal comfort, and energy consumption in residential buildings based on the stacking ensemble learning algorithm
- 10 September 2024
- Published by Springer Nature in Building Simulation
- Vol. 17 (11) , 2125-2143
Developing an integrated prediction model for daylighting, thermal comfort, and energy consumption in residential buildings based on the stacking ensemble learning algorithm
- 28 October 2022
- Published by Elsevier in Building and Environment
- Vol. 226
Data-driven prediction and optimization of residential building performance in Singapore considering the impact of climate change
- 28 October 2022
- Published by Elsevier in Building and Environment
- Vol. 226
Data-driven prediction and optimization of residential building performance in Singapore considering the impact of climate change