Artificial intelligence (AI) is creating shockwaves across all aspects of IT. From a smart building perspective, AI is being integrated into many occupant-facing technologies to help improve the safety, health, and functionality of buildings and campuses. Even more important—and more critical to building operations green initiatives and bottom lines—is the integration of AI and machine learning (ML) to monitor and analyze energy and carbon usage.
While “smart” electrical, lighting, and HVAC systems have existed for a while, their intelligence remains largely static and relies on input by system operators. However, with the advent of AI/ML, these technologies can evolve to the next level by identifying consumption patterns at a far more granular level and can more quickly adapt to these changes, lowering electricity and heating/cooling draw to dramatically low levels. So much so that operators of older buildings who opted to pass on early-generation smart building infrastructure technologies may not be able to ignore them.
MIT energy efficiency pilot highlights the power of AI
In September 2023, MIT announced a pilot program that will use AI and ML to further increase energy efficiency for HVAC systems on campus. Their goal was to use AI to analyze collected from multiple sources including IoT temperature sensors, occupancy sensors, and real-time weather forecasts. Once this training data is collected, it can be analyzed by AI over time to predict what temperature targets specific portions of a building should be set to on any given day and time.
As every building is different, specific data points such as baseline occupancies, locations of occupancy clusters, the amount of time it takes to heat/cool specific rooms, and constantly changing weather conditions are being tracked. Over time, AI can learn these patterns and dynamic conditions to forecast the precise HVAC levels required hours or days into the future. Not only will this help to lower the overall energy consumption and carbon usage over time, but it also provides details into future energy costs and HVAC upgrade recommendations to achieve even more efficiency in the future.
What do building owners/operators need to know about AI?
The science behind various forms of artificial intelligence is quite complicated and based on numerous mathematical models, training data, and tailor-made data center infrastructures. But in reality, building owners and operators don’t have to get into the details of how the technology works. Where things may get complicated, however, deals with the training of in-house IT staff to manage the AI backend so that the system is analyzing properly cured data that can be used to make informed decisions. This requires a sound understanding of what data is collected from each IoT system and how it is used by AI. In many cases, additional training is required.
For these systems to function efficiently throughout the lifecycle of the technology, IT operators must get involved in the design and deployment phases of IoT projects and associated AI. That way, operators better understand how and why systems were designed, how to secure them from internal/external threats, and how to tune them appropriately.
Legacy building HVAC, lighting, and other operational technologies typically save energy and carbon output by scheduling when and where these systems are in operation or programmed to sit idle. While these static processes have indeed been shown to increase energy efficiency, there is still so much more efficiency that can be gained.
For example, it’s not uncommon for HVAC systems to be heating or cooling at full capacity even though a corporate event was scheduled and 90% of employees are at an offsite event. In cases like this, AI could use real-time occupancy information to identify this change in behavior and only heat/cool fractions of the building where the remaining 10% are located. While this is an overly simplistic example, it shows just how easy it can be to save 5-20% of HVAC energy usage and maintenance year-over-year if a more intelligent and finely tuned automation system were in place. AI completely removes the human error factor often attributed to excess energy usage through misconfigurations of systems and manual HVAC adjustments.
The bottom line is that AI represents a significant evolution of smart building energy efficiency technologies. While the technologies are not yet ready for commercial use and will likely be expensive for the first few years on the market, now’s the time to start researching and planning. AI will eventually be a game changer when it comes to smart building operational efficiency practices that leapfrog what is available today. For building operators that have yet to move to smart technologies, this is the time to get on board.