Artificial intelligence for smarter, safer buildings

1 Introduction
2 What is AI?
3 AI in buildings
4 AI and energy use in buildings
5 AI and data
6 Related articles on Designing Buildings Wiki

[edit] Introduction

In building services, as in much of life, it is often hard to distinguish between reality and hype. Since 2010, we have grown used to hearing of smart buildings and smart devices without there necessary being any agreed definition of what these mean. Now there is more talk of applying artificial intelligence (AI) to everything from building design to service and maintenance but, again, without always being clear what lies behind the claim.

[edit] What is AI?

AI implies much more than the ability to follow detailed programmes or instructions, however complex. It entails at least some degree of “learning” ability, to take information gathered in one situation or series of instances and apply it in a different scenario.

It ranges from “weak AI”, which focusses on adapting itself to perform very specific tasks more efficiently, to “strong AI” which has much more general application, more akin to a human brain. AI often deploys “machine learning” to recognise things that can’t be deduced by rules alone.

We have all experienced the joys and frustrations of voice activated devices, as increasingly used in buildings. Learning to recognise and interpret voice commands is a classic example of AI in buildings. AI in the form of voice recognition in a lift might be a simplified version of this.

At BSRIA Worldwide Market Intelligence, we decided that this is a good time to look at what AI is already capable of in building services and what we are likely to see soon.

[edit] AI in buildings

It is important to stress that AI in buildings is still at a very early stage, and claims about its use should be viewed with caution. However, progress is rapid and likely to remain so.

Part of the key power of AI lies in its ability to compare a huge range of different scenarios far more quickly and accurately than any human could. This capability comes to the fore in the design and construction phases of a building. With access to enough data, AI can try many different design options and compare costings, likely energy performance and other factors.

AI can also identify common problems, bottlenecks and safety risks encountered during a construction project and recommend solutions.

Once the building is up and running, service and maintenance provides another ideal application for AI. Malfunctioning building equipment can cause disruption, lost revenue and even endanger lives. By tapping into and analysing the stream of data available from more and more HVAC, security and other building services, AI can identify the tell-tale warning signs of failure, enabling predictive maintenance. This both reduces down times and the cost of maintenance

[edit] AI and energy use in buildings

AI has a huge potential role when it comes to solving some key challenges. With buildings accounting for some 40% of energy consumed, they face a twin challenge of both improving energy efficiency and maximising use of green and renewal energy. Accordingly, we are likely to see increased use of energy generated on-site, through solar PV, wind power, combined heat and power or other means.

Complex and continual calculations are needed to decide whether it is best to use energy from onsite sources directly, store it (e.g. in batteries) or export it to the grid. AI can ensure that these decisions are taken in real time, based on up to date information.

The COVID-19 pandemic of 2020 has promoted health and wellbeing in buildings from a “nice to have” to something that will be needed to make people confident that they can return to buildings. There are already wearable devices that can track people’s physical states and communicate information to the building system. The system can then use this to adjust things like temperatures and ventilation rates in different areas. AI also has the potential to identify possible signs of infection in people in the building.

[edit] AI and data

These uses of course carry their own “health warning” in that they raise very real concerns about data privacy and consent. As AI becomes more sophisticated and more human, these concerns are likely to grow.

AI works best when it has access to the maximum amount and variety of data from different sources and helps us understand the complex ways in which building systems interact, in ways that are both intentional and unintentional. AI is likely to increase the impetus towards more converged and integrated building solutions and also towards more connected buildings.

BSRIA’s white paper, available in May 2021, explores these uses and some of the other key ones.

This article originally appeared under the headline, 'Artificial Intelligence: A Game Changer for Smarter, Safer Buildings?' It was published in April 2021.

--BSRIA