As cities push to cut carbon emissions, buildings are under growing pressure to use less energy while still keeping people comfortable. In zero-energy building design, this balance is especially hard to achieve, El.kz reports citing Interesting Engineering.
Most design tools today rely on static simulations, which means they cannot easily show how heat, airflow, and indoor comfort change as a building design evolves. As a result, architects often have to make important decisions without clear, real-time feedback.
This is a particular problem for Task-Ambience air conditioning (TAAC) systems. These systems control the climate around individual work areas separately from the rest of the room, and they are known to save energy once installed. However, until now, designers have had no easy way to test and compare their impact while a building is still being planned.
AI model brings early energy and comfort testing
In a new study, researchers have developed an AI-powered digital twin that allows architects and engineers to test how energy use and indoor comfort will change while a building is still on the drawing board. Instead of relying on static simulations, the system makes it possible to evaluate different design and air-conditioning options in real time, helping teams spot inefficiencies and comfort issues early, Nanowerk reports.
The project was led by Professor Teng at Kanazawa University in collaboration with a scientist from Fushou University in China. Together, they created a rule-based symbolic AI model called VEEM-ZEB, designed specifically for zero-energy buildings that use task-ambience air conditioning. The digital twin can estimate both energy consumption and thermal comfort at the design stage, giving planners a clearer picture of how a building will actually perform before it is ever built.
Rather than treating a building as one single climate zone, the model breaks air conditioning into two parts: the air around individual work areas and the air in the wider room. This makes it possible to measure both comfort and energy use at the same time using standard PMV and PPD comfort indicators. A built-in VR view then shows these results live, so designers can immediately see how changes in layout, occupancy, or settings affect energy use and comfort.
The system can also run about 48,000 different design and operating scenarios using standard parameters. By testing seasonal changes, different numbers of occupants, and how people behave in offices, the researchers showed that the model can reliably identify more efficient and comfortable configurations, giving designers a much clearer basis for choosing the best energy-saving options.
Helping architects compare cooling strategies in zero-energy buildings
What makes this work stand out is that it moves the evaluation of task-ambience air conditioning from the operation phase to the design phase. Instead of waiting for a building to be built and used, designers can now test how different cooling strategies will perform while plans are still being developed. The system uses a three-layer digital twin that combines clear, rule-based AI with an easy-to-understand VR environment. This lets architects and engineers directly compare air-conditioning setups and control strategies early on, using real numbers rather than guesses.
Instead of relying on estimates made after construction, this approach allows designers to see both energy savings and indoor comfort while a building is still being planned. By showing these two factors side by side, the model helps teams understand how their design choices will affect real-world performance.
The researchers expect the tool to be used in everyday architectural practice as a decision-support system for zero-energy buildings, helping designers make smarter choices that balance comfort with energy efficiency from the very start.