Some years ago, I went to an Internet of Things meetup in Dogpatch, which is a formerly industrial part of San Francisco. The company had a big open space in an old industrial building. Large old windows looked out on the bay, and an endless chain and shaft arrangement allowed manually opening an upper row of those windows. There was a skylight, with the same endless chain window opening system. There were ceiling fans. And there was a modern HVAC system.
None of this was under coordinated control. So as the place filled up, it got too stuffy, but not hot. The main HVAC system did not turn on. And nobody was trying to manually operate all those windows and fans.
Now that's when you want intelligent sensing and control. Sense temp, air speed, and direction, humidity, and rain. (No, don't get them from the Internet, as Nest did.) Inside air temp, CO, CO2, and humidity. Basic motion detectors to see if anybody is present.
Now you can use all that ventilation.
As the number of people increases and CO2 levels go up, open some windows a bit on the bay side. Open the skylight windows a bit and start the ceiling fans in low, blowing upward. That will bring the CO2 down but not over-cool the room. As temperature increases, open windows more and increase fan speed. As the sun sets and incoming air becomes colder, slowly close windows to narrow opening to maintain temperature. Use fan speed to adjust CO2. When everyone leaves, and motion detectors show quiet, open windows more, bring fan speeds to high, and purge the room for a few minutes. Then close all windows, stop fans, and let temperature slowly drop to idle normal, maybe 60 degrees.
Hotels often have "smart" HVAC systems, because they have rooms which rapidly go from zero people to huge crowds. They're in a competitive business, so if customers notice the rooms seem "stuffy", it costs them revenue. Schools have the same problem, but lack the competitive motivation to solve it, so they rarely have CO/CO2/humidity sensing and a control system that can use that info properly. Offices tend to beat the problem into the ground by making the air conditioning system work too hard without using outside air when it would help.
You can buy all the controls for this, but you need an HVAC engineer to get them all working right, and there's a long tuning process. That's where AI comes in. Given an explanation of all the components and a picture of the room, and some time to experiment a bit with the effects of various settings, a system should be able to self tune to the environment.
Incidentally, the IoT presentation was mostly a guy from Samsung touting refrigerators with a tablet in the door. The tablet was not connected to the refrigerator in any useful way.
It didn't do inventory or refrigeration control. It was just a message board.
Some years ago, I went to an Internet of Things meetup in Dogpatch, which is a formerly industrial part of San Francisco. The company had a big open space in an old industrial building. Large old windows looked out on the bay, and an endless chain and shaft arrangement allowed manually opening an upper row of those windows. There was a skylight, with the same endless chain window opening system. There were ceiling fans. And there was a modern HVAC system.
None of this was under coordinated control. So as the place filled up, it got too stuffy, but not hot. The main HVAC system did not turn on. And nobody was trying to manually operate all those windows and fans.
Now that's when you want intelligent sensing and control. Sense temp, air speed, and direction, humidity, and rain. (No, don't get them from the Internet, as Nest did.) Inside air temp, CO, CO2, and humidity. Basic motion detectors to see if anybody is present.
Now you can use all that ventilation. As the number of people increases and CO2 levels go up, open some windows a bit on the bay side. Open the skylight windows a bit and start the ceiling fans in low, blowing upward. That will bring the CO2 down but not over-cool the room. As temperature increases, open windows more and increase fan speed. As the sun sets and incoming air becomes colder, slowly close windows to narrow opening to maintain temperature. Use fan speed to adjust CO2. When everyone leaves, and motion detectors show quiet, open windows more, bring fan speeds to high, and purge the room for a few minutes. Then close all windows, stop fans, and let temperature slowly drop to idle normal, maybe 60 degrees.
Hotels often have "smart" HVAC systems, because they have rooms which rapidly go from zero people to huge crowds. They're in a competitive business, so if customers notice the rooms seem "stuffy", it costs them revenue. Schools have the same problem, but lack the competitive motivation to solve it, so they rarely have CO/CO2/humidity sensing and a control system that can use that info properly. Offices tend to beat the problem into the ground by making the air conditioning system work too hard without using outside air when it would help.
You can buy all the controls for this, but you need an HVAC engineer to get them all working right, and there's a long tuning process. That's where AI comes in. Given an explanation of all the components and a picture of the room, and some time to experiment a bit with the effects of various settings, a system should be able to self tune to the environment.
Incidentally, the IoT presentation was mostly a guy from Samsung touting refrigerators with a tablet in the door. The tablet was not connected to the refrigerator in any useful way. It didn't do inventory or refrigeration control. It was just a message board.