HVAC systems are gluttonous. In fact, they use 15% of the world energy supply. The good news is that in the future this number is likely to drop significantly thanks to digital optimization. HVAC systems are complex and behave in non-linear ways. The current technologies used to control them are far from optimal.
This is where Artificial Intelligence comes into action. Instead of a simple formula-based response, the HVAC AI algorithms create numerous scenarios from a large data spectrum and pick the best scenario for the task.
Here is my pick of the most significant existing technologies which I believe will be at scale in the industry within the next 5 to 10 years.
N°1 : Demand response
With the development of naturally variable energy sources (e.g., wind, solar), balancing energy supply and demand has become a much bigger thing. Spot electricity prices give the opportunity to flexible loads, such as HVAC, to reduce significantly their energy bill by consuming at the most favorable prices.
N°2 : Weather forecast
99.99% of HVAC systems in the world do not use nor have access to weather forecast. This upgrade can bring large improvements in both efficiency and comfort but technically it is not as easy as it might sound. Advanced algorithms are required to make it work effectively.
N°3 : Building behavior and usage
This is one of the biggest issues of existing HVAC controls: they are generic whichever the application and they do not adapt to the specific building and its usage. If on a Friday at 5pm everybody has left the building, why does the control time clock turn-off the HVAC at 9pm?
N°4 : Electrical signal analysis
There is an enormous amount of information contained in the electrical signal of HVAC systems: compressor wear, cleanliness of heat exchanger, fan bearing state… But this is only accessible at a reasonable cost when using energy disaggregation, a branch of Artificial Intelligence, powered by edge computing which brings processing closer to the data source. This is maybe the most futuristic technology of this list, but it holds a large saving potential in HVAC operating cost by predicting and preventing breakdowns.
On the paper, Artificial Intelligence has everything it needs for a roll-out in the HVAC world to save billions:
- Day by day, AI becomes more and more accessible to developers thanks to the continuous flow of open source contributions.
- Electronics and networks have never been so affordable which is critical for data acquisition and control.
But when looking closer, there is a missing bit for a massive HVAC AI roll-out: no one has sorted out the IT infrastructure. Like having high-speed trains but 19th century railroad network. This is what WattSense intends to fix: https://wattsense.com/