Enhancing Thermal Comfort for Building Occupants with a Reduced Carbon Footprint through HVAC Algorithms

Home Research Enhancing Thermal Comfort for Building Occupants with a Reduced Carbon Footprint through HVAC Algorithms
AI algorithm for HVAC

As businesses strive to diminish their energy consumption and curb associated carbon emissions, one particular area ripe for optimization is indoor heating and cooling. HVAC, which encompasses heating, ventilation, and air conditioning systems, typically accounts for approximately 40% of a building’s total energy usage. Implementing methods that conserve electricity while ensuring a comfortable indoor environment for occupants can play a pivotal role in combating climate change.


Researchers at Osaka University showcased substantial energy savings through the implementation of an AI-driven algorithm designed to control HVAC systems. Remarkably, this approach does not necessitate intricate physics modeling or in-depth prior knowledge about the specific building in question. The findings of their work have been published in the journal Applied Energy.


During colder weather, traditional sensor-based systems often struggle to determine the appropriate time to deactivate heating systems. This challenge arises due to thermal interference from various sources, such as lighting, equipment, or even the heat generated by occupants themselves. Consequently, energy is frequently wasted as HVAC systems are activated unnecessarily.


To surmount these hurdles, the researchers adopted a control algorithm capable of predicting the thermodynamic response of a building based on collected data.


This approach proves more effective than attempting to explicitly calculate the impact of a multitude of complex factors that influence temperature, such as insulation and heat generation. Hence, with sufficient data, “data-driven” methodologies can often surpass even sophisticated models. In this instance, the HVAC control system was engineered to “learn” the symbolic relationships between various variables, including power consumption, utilizing a substantial dataset.


The algorithm excelled at conserving energy while still ensuring the comfort of building occupants. Lead author Dafang Zhao states, “Our autonomous system demonstrated substantial energy savings, exceeding 30% for office buildings, by harnessing the predictive capabilities of machine learning to optimize HVAC operation times. Notably, even during the winter, the rooms remained comfortably warm.”


The algorithm’s primary objective was to minimize overall energy consumption, the disparity between actual and desired room temperatures, and fluctuations in power output during peak demand. Senior author Ittetsu Taniguchi adds, “Our system can be easily tailored to prioritize either energy conservation or temperature accuracy, depending on the specific requirements of the situation.”


To collectively work towards the objective of a carbon-neutral economy, it is highly likely that corporations will need to lead the way in innovation. The researchers highlight that their approach could rapidly gain traction, particularly during periods of escalating energy costs, thereby benefiting both the environment and the financial viability of companies.