If you’re the owner or operator of a commercial building, chances are you have invested in a building automation system (BAS) or building energy management system (BEMS). Good move. Buildings are a ripe target for efficiency improvements, and with efficiency, comes cost savings.
In fact, commercial buildings produce around one-fifth of total carbon emissions, with HVAC systems accounting for 40% of buildings’ energy consumption, according to the U.S. Department of Energy.
With the advent of advanced computing hardware and analytics, building owners can leverage artificial intelligence to optimize HVAC operations. By bringing autonomous AI to the built environment, your existing HVAC system becomes a predictive brain that learns precisely how to use less energy to optimize comfort in all zones, at all times.
AI-based BAS and BEMS solutions are already being adopted worldwide. For example, ABB Ability BE Sustainable with Efficiency AI currently manages more than 275 buildings, totaling more than 100 million square feet (disclosure: ABB is my current employer). Collectively, these installations save more than 1 million metric tons of CO2 each year, all by leveraging building automation investments already made.
Vast potential for AI applications
While older buildings have more upside potential in savings, modern buildings have more technology in place that allows for a finer degree of control. As a result, applying AI to any building will likely to produce results. The potential is huge: up to a 25% reduction in energy cost, up to a 40% reduction in carbon footprint, and up to a 50% extension of asset life, according to ABB partner Brainbox AI. For new construction, intelligent HVAC also offers a way to meet energy-related regulatory mandates.
The goal is to make HVAC systems self-correcting rather than predictive while preserving all existing BAS and BEMS functionality. Virtual metering, for example, allows building operators to track energy use at the device level—without physical hardware—by capturing data elements like humidity levels, supply and return air speed and temperature, and current thermostat temperature and set points from the BEMS. Add to that the ability to overlay external data, like weather forecasts, and you have enough data for AI to not only manage HVAC performance, but also alert operators of potential problems when anomalies appear—and before failures occur.
Continuous learning means that AI can adjust digital models in real time, say, after new windows are installed. The building is always “understood” in its current form and can be optimized accordingly. The result is a capability that did not exist even a few years ago. Today’s AI solutions can predict the temperature in HVAC zones up to two hours in advance with 98% confidence. As it learns about the HVAC system and its operations, the system becomes self-healing; that is, it can address problems without human intervention.
First steps to incorporate AI
To begin utilizing AI in a building, an owner typically hires a system integrator to survey existing building systems and assets. Are HVAC drawings available? Is a BEMS present? What does the building’s occupancy look like in terms of numbers of people and duration in a given location? These and other questions will allow the provider to evaluate if an AI solution is applicable to the owner’s building.
From a technology standpoint, all building owners need is networked HVAC controls using an open protocol. The AI will then begin to suggest changes to HVAC operations, testing them first in a virtual environment before deploying them to the live system. The supplier will typically monitor the AI’s progress and perform sanity checks on proposed changes to the HVAC operating algorithm. The AI will look for ways to optimize HVAC assets based on how the building is used and how that use changes over time. These tools also provide data for KPI reporting and supplier experts to notify of potential issues. Depending on building size, owners can expect AI-enhanced HVAC control to produce a return on investment (ROI) within two to four months after the system has learned the building and its HVAC system.
The proliferation of analytic tools made possible by computing advancement and applications tailored for the needs of commercial buildings has brought sophisticated AI-based HVAC control within reach of almost any commercial building larger than 5,000 square feet. We are still in the early days of AI applications in the built environment, but with their compelling business case in terms of ROI and emissions reductions, these solutions will likely become commonplace in new builds and retrofits.
