Use of artificial intelligence in the field of sustainable architecture: Current knowledge

Veronika Krausková, Henrich Pifko

SUMMARY

There is a trend that artificial intelligence (AI) is a direction to take in various scientific disciplines. The idea of ​​AI originated before the 1960s, and with the development of computer technology, the capacities of AI have multiplied. It currently affects many areas of everyday life where we may not even realize that we are already using AI. We are certain that AI is a challenge in the field of architecture and the entire construction industry, where sustainability is one of the current issues.

A notable technological shift in the field of building design is BIM. The participants in the process of design are effectively informed about the current state of the project, but the BIM model should be used for further actions, such as utilising information as interactive tool in the construction, operation or renovation phase.

The objective of the study is to acquire knowledge about AI usability in the optimization of sustainable design BIM processes in architecture. The aim of the study is knowledge of the applicability of AI in the conceptual solution for reducing the carbon footprint in the BIM model of the building of Faculty of Architecture and Design STU. An assumption for formulating the hypothesis is to use AI to predict the pattern of users’ behaviour. Due to need in the older buildings, where massive refurbishment is not possible or appropriate because of the historical or cultural value. So, it is nearly impossible to meet current requirements for energy-efficient buildings.

The article provides a brief overview of AI usability. The base knowledge is presented with terminology in the field of information technology, the processes of artificial intelligence and their applicability to the field of architecture. The context of historical development refers to the considerations of Stanislas Chaillou, who evaluates the connection between the work of an architect and AI as a logical step in technical development.

As examples and sources of information studies of application AI in architecture and building industry were used.

  1. The study by Stanislas Chaillou is about architecture and AI, where it focuses on the ways and processes of using GAN (Generative Adversarial Network) in floor plan design.
  2. The optimization of design processes in BIM is presented by a study that tries to use BIM and artificial intelligence in the design process of buildings. It creates a project called BIMBOT, which generates solutions based on defined priorities for a specific project.
  3. The most discussed part of the article is the work of Ekaterina Petrova. Aim of the study is: Integrating knowledge discovery and semantic data modelling for support of evidence-based design decisions. Suggesting a solution to systematization of building design using the BIM-based design, she draws attention to BIM building models, which contain a lot of information that will eventually remain unused.

Petrova develops the architecture of a comprehensive software “consultant” system using AI and the cooperation of experts in the field of information technology and the building industry. The result of the study will be a “consultant” which collects information from real buildings throughout their life cycle into a robust database. AI methods are used for re-evaluation and sorting of information to gain new knowledge for sustainable building design processes. The main goal of Petrova research is to transfer the acquired knowledge from real constructions back to the design process, in order to create a connection between relevant factual and data-based knowledge within the initial phase of the design of new buildings.

In her work, Petrova also dealt with obtaining specific data from case studies. Based on the statistics, she processed a database of collected information. The data was taken from measurements of two case studies of erected buildings, where she had to find a system categorizing information.

The issue of sustainable building designing with AI in the study is proposed from the perspective of the functionality of information technology rather than from the architect’s point of view. The study addresses the challenges of data collection, processing of information in databases and its compatibility. However, this is the point of utilising knowledge, because it clarifies the idea of ​​correct data and outputs when working with a database which consists of many BIM models.

A dramatic science fiction for which AI is often taken, may be sometimes true. The article outlines challenges and directions of AI development. Some of them are still unrealistic in the current state of scientific knowledge. However, working with AI itself is advancing in various scientific disciplines, such as medicine and disease diagnosis. AI is currently able to find connections in the data based on statistics of the occurrence of a certain phenomenon. AI is also able to recognize that phenomena. To decide based on abstraction, whether to put phenomena in context, AI is still in its infancy and needs the help of a human expert.

AI and humans acquire skills through experience. Some knowledge is very difficult to transform so that AI can “understand” it. If we use AI as a chatbot, it is able to learn concepts based on experience, but does not fully understand the meaning. These issues of working with AI are also transferred to the use in the construction industry or architecture. The article outlines some of these possible problems. Few of them have already been named in studies as specific problems for that case. Other problems are of a general nature or related to the overall development of work with AI. Finally, the study evaluates the potential of working with AI in architecture and ideas about future research.

Keywords: architecture, sustainability, artificial intelligence, AI, building information model, BIM, generative adversarial network, GAN