HVAC – heating, ventilation, and air conditioning – can account for a huge amount of energy usage of a building, whether it’s residential or industrial. Often it’s the majority energy consumer, especially in places with extreme climates or for things like data centers where cooling is a large design consideration. When problems arise with these complex systems, they can go undiagnosed for a time and additionally be difficult to fix, leading to even more energy losses until repairs are complete. With the growing availability of platforms that can run capable artificial intelligences, [kutluhan_aktar] is working towards a system that can automatically diagnose potential issues and help humans get a handle on repairs faster.
The prototype system is designed for hydronic (water-based) systems and uses two separate artificial intelligences, one to analyze thermal imagery of the system and look for problems like leaks, hot spots, or blockages, and the other to listen for anomalous sounds especially relating to the behavior of cooling fans. For the first, a CNC-like machine was built to move a thermal camera around a custom-built model HVAC system and report its images back to a central system where they can be analyzed for anomalies. The second system which analyses audio runs its artificial intelligence on a XIAO ESP32C6 and listens to the cooling fans running in the model.
One problem that had to be tackled before any of this could be completed was actually building an open-source dataset to train the AI on. That’s part of the reason for the HVAC model in this project; being able to create problems to train the computer to detect before rolling it out to a larger system. The project’s code and training models can be found on its GitHub page. It seems to be a fairly robust solution to this problem, though, and we’ll be looking forward to future versions running on larger systems. Not everyone has a hydronic HVAC system, though. As heat pumps become more and more popular and capable, you’ll need systems to control those as well.
Heat pumps can reject or collect heat from water as well. Better to move water or glycol around than refrigerant anyway.
I saw a small (7 barrel) brewery that used one to chill the water-glycol reservoir that supplied their jacketed fermentation tanks. It let them “cold crash” beers that required it and chill down beer after fermentation was complete so they didn’t move heat into their cold room when the beer was pumped into the bright tanks to settle out.
During COVID shutdowns I found part time work at a dairy farm where they used heat pumps and jacketed tanks to chill the raw milk. The heat was dumped into a massive 500 gallon (1892.71 liter) hot water tank so all the farm building had hot water for cleaning purposes.
I object to the idea that HVAC problems are somehow difficult to diagnose. A majority of qualified HVAC service personnel have the equivalent of a high-school diploma, at best. The reason that problems go undiagnosed is due to a combination of lack of awareness by the end users of the systems, lack of self-diagnostics by the HVAC system, and ignorance of the service personnel. These systems are closed loop and easy to diagnose if you have much beyond a basic understanding of how they operate and a notebook with the industry standard “rules of thumb” on hand. Anyone is capable of learning how their residential/smaller commercial HVAC system works and servicing it themselves. The tools required might look expensive until you see the maintenance and repair bills. I started out knowing nothing about HVAC, but not wanting to pay the yearly maintenance bills for my furnace, air conditioning, and later heat pump system. I bought the tools required and learned (with a few failures along the way) until I got to the point where I install my own furnaces/AC/heat pumps without a problem, and the end result is higher quality of maintenance and repair than any of my local HVAC service companies could provide. Another major benefit is that I know the moment that something is wrong – There is no waiting until “oh crap, my furnace doesn’t heat anymore” or “oh no, my A/C isn’t cooling the house like it used to” for me. I am not a smart person by any standard – Most people could do this.
And what are theese tools? What is it that you learned? Please do share :D
“With a few failures” can quickly negate any savings.
Respectfully, I have to disagree. At least in part
What I’ve seen, especially in commercial buildings, are HVAC systems that seem deliberately designed to thwart any attempts by a non-specialist to do anything beyond changing the current temperature. Adding nighttime setbacks, changing deadbands, often even just modifying the operating hours, are not something a layperson can do.
Residential systems are a little more accessible in terms of controls and settings.
I do agree with your point that HVAC problems should be easy to diagnose, but as a practical matter, actually solving the issues is harder than it should be.
Installation of (replacement) residential equipment is indeed almost trivial for almost any mechanically inclined person able to buy the tools and actually read and understand the installation manuals up to a point. Wire type selection, sizing, fusing, air duct and water/fuel pipe sizing are more difficult for new installations but doable for those familiar with the concepts and rough charts and tables are available as well as apps and programs. It’s not whether you can install them, it’s whether you can install them to be safe and operate for a long time as intended. Service can be easy if a flashing LED tells you what is wrong (if it is correct and if it is the ROOT CAUSE), but a halfway decent tech will be able to determine the problem in a couple minutes on a gas forced air furnace or boiler without self diagnostics using a multimeter. There are many cases where even experienced techs get stuck on a symptom that is misleading, especially with newer air/ground source heat pumps (especially high end high temp water to water units). It sometimes happens (rarely, thank goodness) that even the engineers that designed and tested these systems can not troubleshoot what is going on even when every scrap of data taken every 4 seconds is available to them (and not just 1 manufacturer)
This idea does not appear practical when related to a small residential heating/cooling system. There usually are enough built in sensors and fault indicators that any competent service person can use to help diagnose a problem. Especially in newer systems.
In an industrial or commercial setting this idea might work. A comparison of “listening” to a motor, compered to a stored trace of how it sounded a year ago might help diagnose a bearing fault. However, regular PM with a vibration meter and an experienced tech will produce the same results. Tachometers on fan blowers could also be incorporated into the mix. They could show any variance in rotation speed which could indicate a coming HVAC failure.
Most HVAC problems are worsened or caused by knuckelheads thinking they are experienced HVAC techs. How can we AI-out that issue?
Maybe by having a simple monitor that is part of the system that calls an HVAC tech before the knucklehead gets involved
The project description itself seems to have been (almost completely) written by AI and some of the images look AI generated too (if not all). It’s completely unreadable.