Steve Rogerson talks with Paul Marshall, CCO of Eseye, about how smart our buildings will become.

When I was young growing up in northern Shropshire, if the house was too cold, we would light a fire. If it was too hot, we’d open the door.

Many years later, around the turn of the century, I was on a work trip in Taipei and staying at an upmarket hotel. The room was much too cold and adjusting the thermostat did nothing. I contacted reception and they said people can’t adjust the temperatures any more in individual rooms; they had put the whole hotel on one setting as it was easier to handle. Their only solution to my problem was to give me an extra blanket.

One hopes we have moved on from that and the recent advances in smart buildings suggest that is the case. But by how much? I discussed that last week with Paul Marshall, CCO at UK IoT firm Eseye (www.eseye.com).

He said there were many factors coming together driving the smart building market, including the comfort of the people working or living in the building, care for the rest of the world by making everything more sustainable, and dynamic and rising energy prices that have put pressure on building owners and tenants to reduce energy use.

Having a building management system that can handle all these factors is not straightforward, especially in a busy office where different people have different requirements about how warm or cool they like to be. And this changes depending on the time of day.

“You have to balance having the heating off at the weekend and yet still be ready for the first person on Monday morning,” said Paul.

To handle these requirements, the building environmental installation has to have both a connectivity part and a machine-learning part, so it can improve over time as it learns the needs of each occupant. As an aside, I can’t be the only person who, when discussing smart buildings, thinks of the Demon Seed novel by Dean Koontz, and the 1977 film of the same name. All people who design intelligent buildings should be forced to read it.

Anyway, back to the real smart buildings. It is possible now to get inputs from the individuals about whether they are too hot or cold, and feed those data into the machine-learning algorithm so it can fine tune when it turns on the heating or air conditioning. The algorithm can also consider the environment outside the building and adjust accordingly.

“So rather than having a thermostat at each desk, they can tell the building what they like and it can have it ready for them when they come in on Monday morning,” said Paul.

This information can also be coordinated across the whole building and inform decisions about whether the HVAC system needs updating and the type of new equipment that needs installing.

“It can be a challenge to match the needs of the people and the environmental and energy goals,” said Paul.

But when you can, you can also predict what people will be wanting and thus can buy the energy at times of day when it is cheaper and store it for use when needed.

“Providing dynamic pricing information to the building can be quite exciting,” said Paul. “There is a lot of clever learning that goes into that.”

A bigger problem is upgrading existing buildings that were designed and built well before today’s technologies were thought of. However, it is not impossible. As Paul pointed out, there may not be the space to, say, install a heat pump, but the old building will have old equipment that probably takes up more space than what you want to install, so there may be a straightforward answer. And modern sensors and machine learning can help even with an old building in determining what equipment is needed and where.

Paul warns though that many of the features we have been talking about will come in second- and third-generation systems, not in the first-generation systems available today. Also, there is the problem that the people who most benefit from an installation are not necessarily the people paying for it – does the building owner want to splash out on a machine-learning building management system so the person sitting in the corner of the office can have the temperature one degree higher?

“We need to look at all stakeholders,” said Paul. “We have local controls, some level of automated controls, but not the full integration. We want people to take information from multiple sources and control multiple pieces of equipment linked with machine learning to get the benefits for all the stakeholders. Once this starts to come together, then everyone can benefit – the owner, tenant, utility provider and so on. We need everything to start coming together.”

I suspect my old house in Shropshire, if it is still standing, now has double glazing and air conditioning, and I hope the hotel in Taipei does not still have to dish out blankets to keep its guests warm, but many buildings still have a long way to go to reach the dream not of Dean Koontz but of everyone who wants a comfortable place to live and work, and protect the planet at the same time.