How AI can improve building automation system implementation

Artificial intelligence (AI) can make building automation systems (BAS) more efficient and improve energy use and knowledge.

By Chris Vavra January 25, 2024
Courtesy: Chris Vavra, CFE Media and Technology

Building automation system (BAS) insights

  • Embracing AI in building automation systems boosts productivity, addresses skills gaps, and enhances tech support capabilities for engineers and technicians.
  • AI-driven digital twins facilitate autonomy in the building industry, solving workflow problems and optimizing systems for real-time management and efficiency.

Embracing artificial intelligence (AI) tools is critical for engineers that are using building automation systems (BAS) in their day-to-day operations. A panel of experts discussed how they’ve used AI in the presentation “The Evolution and Implementation of AI Applications in the BAS industry” at AHR Expo at McCormick Place in Chicago.

AI uses in the modern business world

Jacob Fenley, area market leader at Cochrane Supply, said modern business have modern expectations due to increased and sophisticated demand and most are still trying to get up to speed with those expectations. Workers, he said, can benefit from large language model (LLM) tools such as ChatGPT to get more productivity out of their time.

AI tools also can help bridge the growing skills gap that has been a major challenge for engineers. Fenley said companies can leverage AI to support modern and legacy BAS systems. This also helps BAS contractors and service providers reduce ramp-up and training times for new technicians. BAS distributors and manufacturers can use AI to boost tech support capabilities and response times.

Left to right: Keith Gipson, CEO/Founder at; Jacob Finley, area market leader at Cochrane Supply; Troy Harvey, CEO/Founder at PassiveLogic.

Left to right: Keith Gipson, CEO/Founder at; Jacob Fenley, area market leader at Cochrane Supply; Troy Harvey, CEO/Founder at PassiveLogic. Courtesy: Chris Vavra, CFE Media and Technology

Moving towards full autonomy with digital twins

Troy Harvey, CEO and founder at PassiveLogic, emphasized how his company is moving the building automation industry, built on AI, toward full autonomy.

AI, he said, will be the foundation of buildings with an emphasis on collaborative software workflows, intelligent resource networks and autonomous building platforms. At the center of it all will be AI and digital twins.

Today’s AI systems are based on deep learning, which is trained rather than programmed and is a method of programming using simple repeating functions trained from data using differentiable programming.

In general, that’s not a bad way for AI to learn, but the approach is impractical for a BAS or building management system (BMS) because every building is unique and no one model fits all. There are many different data sets and it’s a slow training data generation.

Digital twins, which are designed to create a digital representation of a physical place with a digital version, can help with the process by providing a generative AI model with the pre-trained knowledge and helping AI inferencing with a cycle of deduction, induction and abduction. All this leads to autonomy.

“All this lets us solve the workflow problem,” Harvey said. “But AI has to do something useful. We all live in part of a building workflow.”

Generative autonomy, Harvey said, can help merge multiple systems into a single-world view that is managed, aggregated, automated, coordinated and optimized in real time.

Digital twins can play a major role in improving building automation systems (BAS).

Digital twins can play a major role in improving building automation systems (BAS). Courtesy: Chris Vavra, CFE Media and Technology

Using AI to improve opportunity and efficiency

Keith Gipson, CEO/ and founder at, said AI applications in the BAS industry can help building owners and managers because buildings are the largest energy users and 30% of the energy consumed is wasted. Many buildings also lack connectivity and the building and refrigeration controls are not operating at their peak efficiency potential.

AI and machine learning (ML) can help companies optimize their HVAC systems by combining that with proportional-integral-derivative (PID) tuning. Within AI, building managers can use iterative learning control, which uses computational algorithms and ML modeling techniques to perform a control task in repetitive processes, which makes it ideal for PID loop tuning.

This is part of a larger effort to make buildings more efficient and effective in their energy usage and ensuring all energy use is cataloged and understood.

“What we can do is have digital maturity and apply modern technology to the situation,” Gipson said. “Everything we’re doing in building controls is wrong. Why not use the building infrastructure to cool the building?”

Chris Vavra, web content manager, CFE Media and Technology,

Author Bio: Chris Vavra is web content manager for CFE Media and Technology.