As commercial building automation evolves and incorporates features such as data collection and artificial intelligence (AI), its potential for saving energy when applied to HVAC systems increases.

Researchers are studying this potential, modeling energy efficiency measures (EEMs), including improved automation, over millions of square feet of commercial building space, gathering a year’s worth of data from buildings where energy management and information systems (EMIS) were installed to help the buildings run better, and even connecting their own EMIS to 10 buildings at a university campus to boost their performance.

The results, experts say, are encouraging.

“Buildings are full of hidden energy savings potential that can be uncovered with the right analysis,” wrote the authors of a study done for the Department of Energy (DOE) by the Lawrence Berkeley National Laboratory. “With sophisticated analytic software applied to everyday building operations, building owners are using their data to their advantage and realizing cost savings through improved energy management.”

And with recent technological advances, the timing is right, said Grant Salmon, business development manager for commercial controls at Honeywell, which recently launched Forge for Buildings, a program of software, hardware, and services that the company says will help owners and managers make buildings more efficient and sustainable.

“AI and machine learning are hitting a level of maturity that’s allowing (users) to take advantage” and improve building controls, Salmon said.

A Building Tuneup

In a 2017 study done for the U.S. Department of Energy (DOE) by the Pacific Northwest National Laboratory, researchers simulated more than 30 EEMs, including controls- and automation-related measures, over 14 types of commercial buildings. The idea was to estimate the potential energy savings that could be achieved by “re-tuning” commercial buildings using the DOE’s EnergyPlus building energy-modeling software.

Simulated over all 14 building types, the study said, the EEMs resulted in estimated combined energy (electricity and natural gas) savings of 29%. Six controls-related EEMS showed the potential for more than 2% of site energy savings: wider deadbands and night setback (7.8%), shortened HVAC schedules (7.1%), demand-controlled ventilation (DCV) (7.1%), reduced minimum variable air volume (VAV) box terminal damper flow (6.5%), optimal start (5.9%), and supply air temperature reset (2.5%).

Advanced rooftop unit fan controls saved the most electricity, 4% of baseline use, of any measure, but additional natural gas consumption offset that, bringing the overall energy savings to 1.3%, the study said.

Extrapolated, the types of buildings modeled represented 51% of U.S. commercial building space and 57% of the commercial building sector’s energy consumption, the study said.

‘More Controls, Not Less’

Salmon said he’s skeptical when he hears promises of energy savings of more than 20%. But, he added, he’s seen poorly performing buildings in which a 29% savings estimate could be low.

“My gut tells me that energy savings of 10%-15% are most likely from a commercial building controls retrofit, at least in my market,” said Salmon, who works in the New York City metropolitan area. Many commercial buildings in his area, he said, have been subject to energy benchmarking and auditing requirements for years, leading to lower baseline energy use. “Where that is the case, I don’t think 29% savings are typically realistic,” Salmon said.

However, he said, improved automation technologies like fault detection and diagnostics (FDD), adaptive machine-learning software, and equipment-level energy metering can “dig deeper for energy savings” than was possible when the study was published six years ago.

“The future will certainly have more controls, not less,” Salmon said.

Applying Analytics

In the Berkeley study, also from 2017, 15 organizations, with more than 400 buildings and 38 million square feet of space between them, reported savings of 400 million Btu and $9 million energy costs over a year of using EMIS software, which included both energy information and FDD systems. The median whole-building energy savings (all fuels) was 5%, the median , the median cost savings was 20 cents per square foot, and the savings in individual buildings ranged from minus 1.5% to nearly 32%, the study said.

“Participants have made the business case to install analytics, often without outside funding and incentives, because it makes good financial sense,” the authors wrote.

Conservation measures taken through the use of EMIS included optimized HVAC scheduling, reduced overventilation, reduced simultaneous heating and cooling, temperature setpoint adjustments, reduced VAV box minimum flow, and supply air temperature reset. The study noted that some of the energy savings may have come from building improvements not related to the use of EMIS.

In a 2022 study, also from the Berkeley laboratory, cloud-based EMIS software was deployed at 10 buildings, totaling 280,000 square feet, at a university in California, and its impact on energy use studied over about two years. One building, a music building, was all-electric; the other nine were heated by gas-fueled boilers and cooled by chillers. Most of the buildings had older BAS equipment using Building Automation and Control Network over Internet Protocol (BACnet/IP).

Conventional BAS “fail to optimize energy use because predetermined settings become rapidly obsolete, and most BAS are not able to continuously and automatically optimize set points for key systems, such as variable frequency drives, valve positions, and damper positions,” the researchers wrote. “Conventional BAS also do not respond dynamically to changes in building schedules and room occupancy, and they are not cognizant of current and forecasted environmental factors and grid conditions that would enable optimization of energy use.”

The researchers’ software included FDD and a data analytics program they called advanced system optimization (ASO). The platform was designed to provide information about faulty and underperforming equipment and control and optimize HVAC equipment.

Energy savings of up to 35% were achieved by applying FDD, and savings of up to 25% were achieved by applying ASO, the study said.

In one example discussed in the study, FDD analyzed data trends to flag a 120-ton chiller that had been short cycling since its installation in 2016. Once the chiller was fixed, its electricity demand dropped from close to 20 kW to under 15kW and, at times, under 10kW, the study said.

“As HVAC technologies change over time, as they certainly will, we’ll see new opportunities to apply controls in different ways, in better ways,” Salmon said.