Seems there’s lots of companies out there trying to establish credibility by “demystifying” AI?
But it takes more than just talk
It takes real and verifiable energy savings. Off the electric bill!
You can’t fake that.
They should also be able to give at least an overview of and ultimately demo their solution LIVE, like we do. No “deep fakes” here!
There are many approaches to achieve autonomous control of buildings
This is how facil.ai Adaptive Optimization works:
To start you have the system that you’re trying to control. It could be a Full Self Driving (FSD) car or an autonomous building.
With a car it’s all about controlling its speed and position
With any system you have Inputs and Outputs (I/O)
Inputs on a car would be represented by the variable (u).
Steering wheel, accelerator and brake pedals.
The control outputs or states are represented by the variable (x). If the system was just a pure closed loop, with no external disturbances, then it would be pretty easy to achieve stable and automatic control.
Unfortunately there are external disturbances (d) such as other cars, pedestrians, hills, obstacles, wind, etc. This is a non-linear, high dimensional and unknown dynamics modeling problem
Similarly, buildings are among the most complex machines ever built.
What would the system of control for a Central Plant look like?
Well with a Central Plant you have several “systems” of control:
Condenser Water Temperature
Chilled Water Supply Temp
Chilled Water flow/Differential PSI
Thermal Energy Storage (TES)
These and more can be dramatically improved by using AI and Machine Learning
