lethanhluan
publications
/ optimization in architecture design /
/ AI in architecture /
/building energy, simulation /
/ ... /
[7] From hand sketches to daylight performance: a mixed-input neural prediction framework
Le Thanh Luan, Heegun Chong, Le Minh Binh, H. Nguyen-Xuan, Sung-Ah Kim
JOURNAL OF BUILDING PERFORMANCE SIMULATION
[Q1]
IF 2.2
Deep learning can accelerate daylight analysis, but existing methods require multiple tools and complex coding. This paper proposes a streamlined framework enabling daylight predictions from architectural hand-sketches with real-time 3D visualization. The method is implemented based on three main modules: (1) hand-sketch recognition and conversion, (2) mixed-input neural network (MINN), and (3) mixed-input pix2pix (MIpix2pix). Three modules were integrated into a concept application, allowing a comprehensive daylight prediction from a hand-sketched floor plan. Training data were generated using Rhino, Grasshopper, PlanFinder, Ladybug, and Honeybee. The MINN achieved a coefficient of determination above 0.92 for spatial daylight autonomy and 0.959 for annual sunlight exposure. The MIpix2pix2 model generate useful daylight illuminance images with SSIM values exceeding 0.93, closely aligning with simulations. This high-accuracy, fully integrated approach streamlines daylight analysis from concept to evaluation. By simplifying AI-based predictions, the framework offers a practical, efficient alternative to existing workflows.
"
[6] A Physical-Digital Integration Framework for Environmental Simulation through Deep Learning:
Wind Flow Implementation
Le Thanh Luan, Heegun Chong, Sung-Ah Kim
Journal
BUILDING AND ENVIRONMENT
[Q1]
IF 7.1
CITE SCORE 12.5
"
This research introduces a novel four-layer framework that bridges the gap between design with physical models and real-time environmental analysis in architecture. While physical models remain essential for spatial comprehension and tactile design exploration, their disconnect from environmental performance assessment limits their utility in sustainable architecture. Our framework addresses this challenge through four integrated layers: (1) a physical layer for tangible model manipulation, (2) a digital layer for real-time spatial recognition, (3) an AI processing layer for environmental simulation, and (4) an interaction layer for visualization and control. We demonstrate this framework through wind flow analysis implementation, developing a multimodal pix2pix model that achieves wind flow prediction with SSIM values of 0.754 and PSNR of 22.630, trained on 603 apartment complexes across five South Korean cities. The digital layer employs ArUco markers for robust object detection, while the interaction layer integrates the Mixtral-8 × 7b language model for natural parameter control through a web-based interface. Physical prototyping and user evaluation validate the framework's effectiveness, confirming its ability to preserve intuitive design workflows while providing immediate environmental feedback. By integrating physical modeling with real-time analysis, the system demonstrates significant potential for transforming architectural practice, education, and stakeholder engagement, while establishing a foundation for expanded environmental assessment capabilities.
"
[5] Mixed-input neural networks for daylight prediction
Le Thanh Luan, Sung-Ah Kim
Conference
ICCEPM2024
SAPPORO
JAPAN
"
In this research, we present the implementation of a mixed-input neural network for daylight prediction in the architectural design process. This approach harnesses the advantages of both image and numerical inputs to construct a robust neural network model. The hybrid model consists of two branches, each handling in-depth information about the building. Consequently, this model can effectively accommodate a wide range of building layouts, incorporating additional information for enhanced predictions. The building data was created utilizing PlanFinder in Rhino Grasshopper, while simulation data were generated using Honeybee and Ladybug. Weather data were collected from three distinct localities in Vietnam: Ha Noi, Da Nang, and Ho Chi Minh City. The neural network demonstrates outstanding performance, achieving an R-squared (R2) value of 0.95 and the overall percentage difference in the testing dataset ranges from 0 to 20.7%.
"
[4] Hand drawing-based daylight analysis using deep learning and augmented reality
Le Thanh Luan, Nguyen Xuan Hung, Sung-Ah Kim
Journal
RESULTS IN ENGINEERING
[Q1]
IF 6.0
CITE SCORE 5.8
"
The utilization of artificial intelligence (AI) in daylight analysis has recently experienced an increase. Its capabilities have been demonstrated through different studies. While using AI for daytime simulation can reduce time, it also presents significant challenges. Whether we utilize the conventional approach of simulating daylight or employ artificial intelligence, the process of going from hand drawing to obtaining the daylight analysis result involves intricate procedures associated with modeling, programming, and integrating AI models. This paper introduces an efficient and simple approach to rapidly predict daylight performance directly from the user's hand-drawn sketches. The core idea is to leverage a deep learning model for predicting directly from hand drawing and augmented reality (AR) to seamlessly display results in an AR environment, overlaying them on the original drawings. This method not only increases user engagement and accessibility but also eliminates the dependency on traditional daylight simulation software and workflow. Artificial neural network (ANN) was exploited using data derived from 3D building models and Useful Daylight Illuminance simulations created by Rhino3D and Honeybee. The color code method, various imaging, and data processing techniques were implemented to ensure compatibility between the hand-drawn input and the simulation model. Unity3D is used as the development platform with the AR engine Vuforia. Notably, the ANN model yields promising results, with R2 ranging from 0.872 to 0.985. This is achieved by efficiently capturing and converting grid-based hand-drawn images into suitable input for the ANN. The obtained results demonstrate the feasibility and emphasize the concept's practicality.
"
[3] Game-based Platform for Daylight Analysis using Deep Learning
Le Thanh Luan, Sung-Ah Kim
Conference
eCAADe 2023
GRAZ
AUSTRIA
"
Daylight analysis is not easy and requires skills in specific software and techniques andsignificant computation time. These skills are necessary for architecture education, butsome students may find them challenging. For this reason, a software-free andsimulation-free approach that quickly calculates daylight performance may be a moreeffective way for students to learn and practice architecture design. From these ideas, agame environment, which is familiar to the young generation, may enhance the excitementand engagement of education in this field. The development of a cubic builder gameplatform that utilizes the Deep Learning Model (DLM) to help users learn about daylightanalysis within the game environment is currently underway. This paper presents thepreliminary results of this study that focused on exploring methods for implementing andusing DLM to predict daylight performance in a game environment. Using a drawingcanvas, users can give design inputs in this environment. A framework involving threesteps has been developed to combine data from the design and gaming environments.First, small-scale building models with specific design contexts and simulation data werecreated in Rhino and Grasshopper using LadyBugs and HoneyBee. Second, a DLM wastrained on these data to make predictions. Last, developing the game environment withthe well-trained DLM in Unity3D. Through analysis, the DLM's performance in gameenvironments confirmed the potential of this approach. A building system will fullyimplement the game environment in future research. The DLM's predictive performancewill be enhanced using more extensive and diverse data sets.
"
[2] Machine learning-based real-time daylight analysis in buildings
Le Thanh Luan, Nguyen Thi Viet Ha, Jeahong Lee, H. Nguyen-Xuan
JOURNAL of
BUILDING ENGINEERING
[Q1]
IF 7.144
CITE SCORE 6.4
SJR: 1.164
"
Daylight analysis is essential in building design to ensure indoor environment quality, including health and thermal comfort vis-à-vis energy. It is a repeating and time-consuming process of design options. Several studies conducted machine learning models to accurately predict daylight performance in particular design situations. Therefore, developing an AI-based real-time daylight analysis platform becomes more promising. However, buildings can be designed with arbitrary shapes, creating a real challenge for the AI to recognize any building layout. From that perspective, the idea of finding the design variables that characterize all the building layouts becomes the key solution. To unlock this challenge, we promote a novel method of creating design variables and building a machine learning model that can efficiently forecast daylight performance with different building layouts. The daylight metric was Useful Daylight Illuminance with four ranges, and the case studies were assumed medium-sized buildings located in Ho Chi Minh City, Vietnam. All the data for training and predicting were created by the simulation DIVA tool. Obtained results showed the excellent performance of the proposed approach, which brings more promising in developing a data-driven machine learning platform for real-time daylight validation. Moreover, the present framework can adapt to any specific machine learning model or daylight simulation tool and daylight metrics.
"
[1] Optimal design of an Origami-inspired kinetic façade by balancing composite motion optimization for improving daylight performance and energy efficiency
Le Thanh Luan, Duc Thang Le, Hung Minh Ngo, Hung Quoc Nguyen, H. Nguyen-Xuan
JOURNAL
ENERGY
[Q1]
IF 9.0
CITE SCORE 15.3
SJR: 2.17
"
This article presents a novel concept for an Origami-inspired shading device based on dynamic daylight that can be used to improve the daylight performance of a target building and reduce the energy consumption for the building. The daylight performance is evaluated based on the Leed v4 (Leadership in Energy and Environmental Design) daylight criterion. The proposed shading device is experimented in an office located in Ho Chi Minh City, Vietnam, where there is a tropical monsoon climate being hot and humid by the year. To investigate the effectiveness of the proposed design in acting as a sun shading system for the office, we consider eight cases corresponding to eight directions which are South, North, East, West, South-East, North-East, South-West, and North-West. An automatic simulation optimization procedure is developed by combining a daylight simulation tool called DIVA and an optimization method called Balancing Composite Motion Optimization (BCMO). BCMO is used to find the optimal design for the proposed kinetic shading device which will help the building to improve daylight performance. It must be noted that the proposed framework is not necessarily tied to any particular optimization tool or type of building. The results show that the proposed kinetic device has outstanding performance as it helps the building to achieve 2, 3 points in Leed v4 for four different directions, including North, North-East, South, North-West.