High Tech Airflow Control for Smarter Energy Use

Using high tech to reduce energy use, specifically through machine learning software applied to HVAC systems, is not new to commercial real estate. Doing so by moving beyond timing of heating and cooling entire areas in coordination with foot traffic patterns is. Add on an interest in manipulating air currents to also improve general health in a building by flushing out airborne viruses and you have a new approach from Mitsubishi Electric Research Labs.

“Our journey began with doing research on air conditioning and reduction of power consumption while maximizing thermal comfort in buildings and offices,” principal scientist Saleh Nabi tells GlobeSt.com. “If you have a big office and only part of the office is filled with people, why do you have to heat up or cool down the whole place?”

Unlike an airplane, where ventilation can be shut down in various areas, Offices are usually full of open areas that aren’t so partitioned. “You need to find a smart way to control the flow,” Nabi says. “The short answer is we found the way to move heat from one place to another place to save energy. But when Covid happened, we realized we were separating the office into different zones.”

The answer was … math. Strictly speaking, the mathematics of fluid dynamics, which is the study of how fluids like liquids and gasses move in and around objects in the presence of forces that affect the flow. This involves complicated and difficult-to-solve equations that depend on a number of independent variables and then using the answers to direct how air louvers should direct air coming out of vents.

“We’ve solved it for various scenarios that we think are going to happen,” Nabi says. “We have pre-stored solutions we’re waiting for.” Thermal cameras in the vents help detect hot spots and bodies. The software-controlled louvers then redirect the air. Except the doesn’t happen with only one vent. It can involve all the vents in a space because the air currents meet and interact. It’s the final mix of air that Mitsubishi is trying to control.

Eventually, Mitsubishi looks to incorporate machine learning so the system can be used in any given space. That way the HVAC control system could learn how to better control air for a given space.