Chemical engineering researchers study use of AI to mitigate viral transmission in schools

Tuesday, September 24, 2024

Media Contact: Tanner Holubar | Communications Specialist | 405-744-2065 | tanner.holubar@okstate.edu

Traditional heating, ventilating and air-conditioning systems manipulate a building’s environment but do little to stop the spread of airborne contaminants. These can cause great harm, especially to those more vulnerable, such as children.   

College of Engineering, Architecture and Technology researchers are beginning a three-year, interdisciplinary project titled “AI Enhanced Risk Assessment for Mitigating Indoor Viral Transmission in Public Schools,” funded by the National Science Foundation.  

Dr. Yu Feng, associate professor in the School of Chemical Engineering, is working alongside Dr. Chenang Liu, assistant professor in the School of Industrial Engineering and Management. 

Their research combines a computer model that captures how airflow affects viral transport with a generative artificial intelligence model. The model will be trained using their lab's computational fluid dynamics and host cell dynamics simulation data to improve the design and real-time control of air handling technology.   

“The research could reduce infection risks by optimizing HVAC systems to reduce the spread of airborne pathogens, particularly in schools, hospitals and other crowded indoor spaces, “ Feng said. “Especially with the support from Dr. Changjie Cai at OUHSC (Director of the Children’s Environmental Health Center in the U.S. Southern Great Plains), this project has a substantial impact on public health improvements in protecting children from airborne infectious disease.”  

Using state-of-the-art, reliable simulations could help improve energy efficiency while maintaining high air quality and lowering costs. The knowledge gained in this research could help improve public health policy and building design standards. 

For an HVAC system to have the capacity to mitigate airborne contaminants, it would need high-efficiency filtration; increased fresh air ventilation, ensuring higher rates of outdoor air in the system; optimized spatial distributions of airflow velocity, pressure, temperature and humidity.

A properly designed ventilation system has "smart” controls that can optimize airflow in each room to remove contaminants efficiently, as well as not recirculating the air in a harmful way. Real-time monitoring, such as sensors that continuously measure air quality and pathogen levels, can also help mitigate the viral spread of airborne contaminants.   

Artificial intelligence can be vital in improving HVAC design and real-time data. AI models can utilize a large amount of data to suggest the best HVAC configurations and operational conditions based on the room's layout, air quality and infection risk.   

AI can also provide real-time information on airflow in an HVAC, using sensors placed throughout the ventilation system. This can dynamically adjust airflow, ventilation rates and filtration settings based on the most current data.   

Generating infection risk indices 

How airborne virus-laden droplets travel through indoor environments can be simulated by combining computational fluid dynamics and host cell dynamics. The CFD model predicts airflow patterns, droplet transmission and droplet deposits within indoor spaces and people’s respiratory systems.   

It will also monitor air velocity, humidity and temperature and how they affect airborne contaminants. The HCD model will predict a person’s response to inhaling a virus, including virus replication, immune response activation and progression of infection.  

“Those outputs from the CFD-HCD simulations are key metrics for calculating an infection risk index,” Feng said. “Specifically, by evaluating factors such as viral load, deposition in the lungs and host immune response, the derived IRI can be used to assess the safety of different HVAC designs and classroom layouts.

The generative AI model will utilize generative adversarial networks and diffusion models to optimize HVAC designs. It will analyze CFD-HCD simulations and then suggest specific configurations that minimize infection risk while being energy efficient.   

Inspiring the next generation  

There are plans to involve K-12 students through research activities, including workshops and hands-on activities. One established outreach program, which is a part of Grandparents University, is called “Lungevity.”  

K-12 students are introduced to lung health, virus transmission and air quality, and have an interactive experience, including building lung models and using virtual reality to see pulmonary functions and how virus-laden droplets are transported.   

“To maximize the potential impact of the outreach activities, recorded lectures and resources will be made available to students and teachers online,” Feng said.   

Graduate and undergraduate students will run simulations, develop AI models and analyze data as part of the project, as well as help mentor and interact with K-12 students and help with educational outreach.   

A truly interdisciplinary endeavor  

This research project is evident of solid collaboration between experts at OSU and OUHSC. By leveraging cutting-edge AI, CFD and public health expertise to tackle a worldwide challenge.   

The application of this project extends beyond schools, as public health can be improved anywhere with a traditional HVAC system and contributes to a broad vision of healthier and more resilient communities.   

Feng said he is thankful to have collaborated with Liu of IEM and Cai at OUHSC.   

“I believe there are numerous ways still lingering there to enhance pulmonary health for residents in Oklahoma, particularly in rural areas,” Feng said. “This could involve both research and educational initiatives. I am eager to establish new collaborations with researchers to advance these efforts.”    

Learn more about Feng and his research here.