Adding AI-based Controls to Smart Buildings for a more Responsive Built Environment
There are 5 main reasons why buildings today are inefficient
- Buildings may run 24x7; people do not.
- Buildings supporting critical operations should run accurately and error free; the more critical the operation, the less error tolerant it is.
- A human action is relatively more expensive than a machine action.
- Building use can change
- Seasonality can impact operating context
In the past, many of these problems have been targeted through building automation systems.
- Building automation can be set to run at all hours
- Boundary conditions can be set in advance
- Building operations can be controlled centrally by the BMS/BAS
But there are key problems that are unsolvable by the state of building automation today. These problems relate to change and the rate of change.
Some things that change faster and more frequently than a traditional building management system can accommodate include:
- Building operator objectives
- Infrastructure assets in the building
- End use and density of the space
- Seasons and climate
- Utility, rate scheduling, pricing, market incentives
- Occupant behavior
Today, building automation is expensive, slow and difficult, and it’s getting harder because the rate of change is increasing. The built environment is more dynamic than ever before.
In the past 10 years we’ve seen the emergence of new technologies that make ambient intelligence possible.
In just the past decade, the cost of sensors has halved. The cost of connectivity and bandwidth has shrunk by 10x. The cost of compute has dropped by 40x. And the cost of storage has fallen 100x.
We now have all the ingredients we need to add intelligence to building controls.
These smart building ingredients in total are Verdigris Adaptive Automation, which launched just this past week.
We’ve integrated networked energy sensors into the building to collect real time data. We’ve developed software that can recognize your context. It “learns” building behaviors and develops an inference for the operating pattern for it. We’ve created algorithms that are adaptive to your occupants and environment. And we’ve use machine learning to add predictive capabilities to anticipate the future.
These superpowers combined make Verdigris Adaptive Automation-enabled buildings uniquely excellent at adapting to change and predicting the future. In turn this results in more performant building operations and greater financial accessibility.
In our current release of Adaptive Automation, we’re introducing these core capabilities for buildings with centralized control systems: dynamic adjustment of RTU output air temperature, dynamic chilled water output temperature, and dynamic thermostat temperatures.
Where a traditional BAS can react to pre-modeled setpoints that are typically set at building design, it must be re-analyzed, remodeled, and reoptimized periodically to maintain efficiency. Conversely Adaptive Automation can make hundreds of changes per month in response and anticipation of changing operating conditions autonomously.
Our roadmap includes layering additional capabilities to manage battery storage, EV charging, other types of air handling units, exhaust fans, and connected devices all in a seamless, transparent, and responsive way.