[00:00:00] Speaker A: Hello and welcome to the Consulting Specifying Engineer podcast. I'm your host, Anna Steingruber and today we're talking to Richard Vedvik, a senior electrical engineer and acoustics engineer at IMEG Corporation. Today we'll be talking all about artificial intelligence and building automation and how it can reduce energy and reach net zero goals. AI is definitely one of the hottest topics in the industry right now, and engineers and building owners are working to use the new technology to to increase comfort for occupants, reach sustainability goals, and better predict patterns of occupancy and building use.
Now I'd like to officially introduce our guest, Richard Vedvik. Richard has experience in healthcare, education, commercial and government sectors as both an electrical and acoustical engineer. He has helped develop in house software tools and custom acoustics test equipment to assist searching for and solving noise problems and performing field measurements of equipment. He was a 2015 consulting specifying Engineer 40 under 40 winner, and is a member of our editorial advisory board. Welcome and thank you for joining us.
[00:01:05] Speaker B: Thank you.
[00:01:06] Speaker A: Yeah, we're excited to get to talk to you today. So, as we begin, can you explain how buildings are controlled currently?
[00:01:15] Speaker B: Absolutely. I like to think of it as a passive system, a reactive system.
What I mean by that is when you enter a room, the controls aren't going to respond immediately to a change in occupancy.
Now, if it was predictive, it would know within 15 to 20 minutes the room is going to get warmer and I should probably turn the air conditioning on. But with our current building automation systems and current equipment, it waits until a sensor that's located somewhere in the room says that it's hot enough to trigger a cooling threshold and then it asks for cooling. And after everybody leaves that room, the building automation system doesn't really know that. Now, there are some variations in which we might do controlled occupancy using the lighting occupancy sensors, but even those will sometimes have a 15 to 30 minute delay. That means the building is now cooling a room using energy that nobody's in. And it might be a room that nobody's going to come back to at any point in time.
We see that same concept through all of the various systems and all the various sensors within these buildings. The only thing that's predictive is the human interaction. People pressing buttons and developing controls in real time.
Most building automation systems are running on programming. The modern ones anyway, the old pneumatic tube systems or pneumatic air systems, they will react to sensors only, and the programming is only through traditional old school logic.
So we Might have some predictability within the computerized building automation systems, but it certainly isn't adaptive. You write a program with if then statements. If this happens, then do this. And that's the limitation of a lot of the systems that are controlling buildings today.
[00:03:13] Speaker A: So based on all of that, how do you think AI can be utilized to improve the way that buildings are controlled?
[00:03:21] Speaker B: Well, personally, I foresee a future in which buildings, both large and small, commercial, residential, etcetera, Will have active AI to which the world is very slow.
Utilizing access to weather, access to schedules, access to cameras and other analytics, being able to know this room is about to be occupied for half an hour and the recipient acceptance rate is historically about 75%. Therefore, I can predict what that heat map or the additional warmth is going to provide. So instead of letting it get hot and then trying to react to cool it, I will proactively react to that occupancy. And then 15 minutes before the meeting's about to about to end, I can release and shut that system off and give it time to coast back up. And then everybody leaves. And now we've just saved a tremendous amount of energy.
I foresee a future in which AI will be able to manage these building automation systems to their peak efficiency, because they're going to be constantly, constantly looking at it over and over and over again.
That would be a very mundane task for humans to have to undergo. Constantly checking the weather every minute, looking at radar, understanding the outside of the building is going to do to then predictively and preactively act to try to minimize energy usage. Additionally, AI can coordinate across an entire campus. So the overall simultaneous, the peak load of that campus can also be maintained by sharing tasks. You're going to do this and then I'll do something, and then you'll do something that we're going to share the amount of energy that we have available to us instead. It's currently a on demand and we see that in our utility rates for our peaks. And we know that there are currently programs available to building owners to eliminate their energy or reduce the extra excess energy that occurs during these peak times.
Now, there are some traditional methods of controlling it and maybe some manual interactions that a savvy building control system can pull off. But with AI, it would be seamless.
It would all happen essentially in the background.
The other benefit that AI provides is the learning. Language models allow plain conversation. You simply speak to the building automation system and ask it to do something. You're not trying to get into a program through a couple terminal stations and then writing and modifying code, just like we're used to with some of our in home devices where we can ask it to turn lights on, set timers, pick music.
That same logic can apply to buildings. You can simply talk to the system and explain.
Today at 5 o' clock there will be a function in the lower ballroom. There will be approximately 200 people there. Please plan accordingly. This is an alteration to our normal schedule. And at that AI says got it. I understand what you're saying. I normally would turn off the H Vac system, let it coast, but in that space I will make sure that I keep it conditioned and prepare for the occupancy change.
So those are some examples that I have off the top of my head. And I'm sure many of the people listening to this can think of other ways that I can simplify, automate and expedite the building automation system.
[00:07:05] Speaker A: Right. That sounds like it would be a good idea and a useful thing in a lot of buildings. But do you think that all kinds of buildings are ready to implement these kinds of AI systems?
[00:07:17] Speaker B: Unfortunately, I would say not most. I would say most buildings are still running older antiquated type systems that are not yet capable of being fully automated. We know that a lot of buildings still have manual valves and they still have pneumatic controls, things that we can't yet take over.
Now, over time and over the past 20 years, we've seen a shift where new modern building automation systems are being installed and they are being done in a piecemeal format, typically because the budget of that little project can't afford to replace a million square foot of pneumatic and outdated controls.
However, in order to maximize energy savings and to allow AI to take over, not in a terrifying way, but simply in an energy saving way, more expansive projects to upgrade and finally upgrade the rest of the systems will be necessary.
And that investment does have a return.
We know that the energy savings that we're going to experience will certainly allow for a payback period that is probably fairly rapid.
[00:08:37] Speaker A: And so do you think that the national initiatives for net zero by specific deadlines, whether that's 2030, 2040, are those helping this transition towards using AI systems in these buildings?
[00:08:50] Speaker B: Yes, I do.
I believe that's the motivation for a lot of facilities to start investing into these systems. We know that it's going to cost money. We know that making these commitments does not come free. However, we do expect a return on that investment.
We expect that not only through energy savings, but through a simplification of how These systems are monitored and controlled.
We're not looking to replace a current jobs because I don't see enough people in those jobs currently.
What we're looking to do is make sure that the people that are in those positions can work effectively. We want to help them, we're their partners and we want to make sure that they can do the job as best as possible, knowing that that's their intent and that's what they're trying to do. Now, the biggest hindrance typically is finance. Those infrastructure projects don't have an obvious return on investment because they don't have an obvious payback with services rendered.
However, from an energy saving standpoint, they certainly can pay for themselves.
[00:09:58] Speaker A: But it sounds like you think that the main benefit of AI in building automation systems is going to be in energy efficiency. Do you think that there are other maybe opportunities that AI can help in, or is that the biggest and most major one that you would like to touch on?
[00:10:13] Speaker B: I understand that there's hesitancy within the the movie creation of what we think AI having control and access could provide. We've created as a population a fear around giving control to the robots, so to speak.
However, it's still a program that does react based on a series of limited inputs. And I believe that we can leverage the mundane, repeatable tasks to computers that can execute them fairly rapidly and with good result.
Now, I don't believe that it replaces everything though, because it writes fairly poorly and it communicates in an awkward, almost predictable way. And I believe that humans are savvy enough to learn how to work with this new technology just the same as we've learned to work with cell phones and Zoom meetings and all the other advances that we've seen in the last 20 years.
I look forward to the next generation of designers, engineers and facility staff to look to see how they can be creative with AI leveraging that technology to simply get more done in a day and to manage energy usage in the most environmentally responsible methods possible.
[00:11:39] Speaker A: Yeah, great. Is there anything else that you didn't touch on at all that you want to say?
[00:11:46] Speaker B: Hopefully, as the push towards 2030 and 2050 goals for net zero come into fruition, my hope is that we work together to achieve those goals.
I hope that we don't see infighting or competition. This is going to be a team effort. It will require releasing a little bit of, I would say, our vulnerabilities because I need all buildings in a campus or on a micro grid to be able to communicate with each other. So the individual AI programs within each of those buildings can coordinate energy usage for each of those micro grids and also with the utility. So hopefully we don't see too much privatization of that technology in which it's being used to generate revenue, but instead we have the more altruistic goal of trying to make the world a better place overall and working together as a team.
[00:12:47] Speaker A: Yeah, that's great. Thank you so much for joining and for speaking today.
[00:12:51] Speaker B: Thank you very much for the opportunity.
[00:12:54] Speaker A: Yeah, of course. So this has been Richard Vedvik on AI and building automation and as AI becomes more and more commonplace, conversations like these ones are necessary. Its impacts can be seen across industries and it can do a lot of jobs that humans aren't capable of, which can help engineers and building owners in the long run. For more information on AI, please visit consulting specifying
[email protected]. don't forget to check in every other Tuesday to hear more from industry experts. Thanks for listening and we'll talk next time.
