Research is increasingly based on data and on how data can be used to arrive at better designs, decisions or measurements. It sometimes requires enormous computing power to process all models and scenarios. Soon, all researchers, students and educations at TU Delft will be able to use the computing power from DelftBlue to solve and calculate complex problems. This high-performance computer has a speed of no less than 2 petaflops (one million times a billion calculations per second). This supercomputer can also be used by the Faculty of Architecture from the end of April.
High-performance computing at BK – Delft AI Labs
Architecture is also working on innovations that can be made possible with the supercomputer. Examples can be found in the three Artificial Intelligence (AI) Labs started at BK as part of the TUD-wide AI Labs program. In these AI Labs, experts ‘In AI’ work with experts from different domains to promote AI within these applications. The three AI Labs at Architecture are:
3D Urban Understanding: (3DUU)
Thanks to developments in photogrammetry (collection of information from photos with emphasis on dimensions) and 3D computer vision, we can now capture the built environment in 3D on a large scale. In 3DUU Lab, new methods and techniques are developed to automatically recognize and model objects in the built environment in 3D using AI. The techniques will automatically detect what each object or surface in the data represents, such as a building, a tree, a traffic light or a terrain plan. This data can be used in various applications, such as planning, management, visualization, analysis and simulation. Based on this information, we can gain insight into the current or future state of a particular area of an city to make the best decisions about planning, design and management in urban areas.
Artificial intelligence for design, analysis and optimization in architecture and the built environment: (AiDAPT)
At AiDAPT, AI methods are being developed to help architects and engineers analyze large amounts of available data (design, image, video, measurements in and on buildings and structures) to further optimally integrate this data into their complex decision-making process. From initial design to life cycle planning, the creation and operation of the built environment involves complex decisions and analyzes. This complexity exceeds the capacity of existing computer methods to process available data and optimize decisions on a large scale, giving us unprecedented opportunities to better support and guide our decisions. The complexity of architectural and technical applications provides a fruitful basis for the validation and development of AI methods for domain-specific intelligence.
Activating Intelligence in Building Sustainable and Viable Environments (AiBLE Lab), Leader: Dr. Tong Wang, MBE (and EEMCS)
The built environment faces major social challenges. Planners, architects, and citizens who use or live in these spaces can all have different views and priorities. Any strategy to deal with the challenges of the built environment must take into account different preferences during the design process. AiBLE Lab explores how artificial intelligence can be developed and used in complex real-world scenarios. The goal is to help reach effective, transparent and sustainable decisions and agreements. This means incorporating human feedback into the process, continuously improving decision-making and driving behavior change. The laboratory focuses on two important challenges in the built environment: energy conversion and circularity.
Education at BK can not do without high-performance computing, such as Computational Design Studio BK7083, and the two electives offered by MSc Geomatics: Machine Learning for the Built Environment GEO5017, and Modeling of Wind and Propagation in Urban Environments GEO5015.
With the development of the DelftBlue supercomputer, the Delft Institute for Computational Science and Engineering (DCSE) meets the enormous needs of researchers and students for advanced computer resources.
The goal of DCSE is to bring all TUD researchers in computational science and engineering into contact with each other and to intensify research across faculties. The education offered by DCSE is aimed at bachelor’s, master’s, PhD students, postdocs and all staff. The department has existed since 2017. Another five-year period was awarded at the end of 2021 and is supported by the six founding faculties and the Faculty of Architecture, which have recently joined.
You can request access to DelftBlue via the self-service portal.
- For architecture, Professor Jantien Stoter is the contact person for both DCSE and DelftBlue. If you have questions about this, you can contact her.
- 3D Urban Understanding: (3DUU), Leader: Dr Liangliang Nan, Urbanism (and EEMCS)
- Artificial intelligence for design, analysis and optimization in architecture and the built environment (AiDAPT). leaders: Dr. Seyran Khademi (architecture) and dr. Charalampos Andriotis (AE&T)
- Activating Intelligence in Building Sustainable and Viable Environments (AiBLE Lab), Leader: Dr. Tong Wang, MBE (and EEMCS)
If you want to know more about DCSE and DelftBlue, visit DHCP’s website, where you can also sign up for DCSE’s newsletter or watch the video TU Delft | Delft High-Performance Computing Center (DHPC):
Previously published news: TU Delft supercomputer officially opened
Image caption: photo by JJ Ying via Unsplash.