Machine learning helps create “autonomous” building

Autonomous vehicles have rapidly become a buzzword in the electronics industry. But autonomous buildings that can learn and regulate their own energy usage may not be far behind. Honeywell has developed a cloud-based, closed-loop, machine learning solution that continuously studies a building’s energy consumption and automatically optimizes energy settings without compromising occupant comfort levels.

Honeywell terms its solution Honeywell Forge Energy Optimization, which the company claims can deliver double-digit energy savings and be implemented without significant up-front capital expenses or changes to a building’s current operational processes. The company tested the solution during a pilot at Hamdan Bin Mohammed Smart University (HBMSU) in Dubai, United Arab Emirates, demonstrating an initial 10% energy savings. HBMSU is the first accredited smart university in the UAE and is known for its technology and innovation programs.

The Forge solution was applied to HBMSU’s existing building management system, which uses a competitor’s technology to demonstrate the platform’s open architecture and hardware-agnostic capabilities. The additional energy savings is considered significant, as HBMSU is already regarded as an energy-efficient building with fully connected lighting, cooling, building management, power and efficiency control that is optimized based on real-time occupancy. The pilot also uncovered local control issues with the chiller plant and fresh air handling unit that were not adjusting to set points.

“As a smart university, we look to deploy the latest technology across our campus and ensure our buildings are efficient. We were pleasantly surprised by the results we saw from Honeywell Forge and its ability to drive further energy savings beyond our achievable optimization with the techniques we have,” said Dr. Mansoor Al Awar, Chancellor of Hamdan Bin Mohammed Smart University, in a statement. “Our further partnership with Honeywell will help to support the advancement of artificial intelligence (AI) modeling for building automation and provide our students with first-hand applications of how AI and machine learning (ML) will drive operational efficiencies in buildings.”

According to a company spokesperson, the Forge solution uses a variety of environmental monitoring sensors, some of which come from Honeywell. As the systems integrator, Honeywell can integrate sensors and other components from various suppliers depending on which work best, the spokesperson said in an e-mail to FierceElectronics.

Honeywell Forge Energy Optimization continually optimizes a building’s internal set points across hundreds of assets every 15 minutes to evaluate whether a building’s HVAC system is running at peak efficiency. When Forge finds a need to make an adjustment, it analyzes factors such as time of day, weather, occupancy levels, and dozens of other data points to determine the building’s optimal settings and makes calculated decisions 96 times per 24-hour period for every building in a portfolio, 365 days a year.

According to Honeywell, Forge Energy Optimization is simple for building portfolio owners to deploy with plug-and-play capabilities. Implementing the system does not requires major changes to the building’s infrastructure.