(313c) Real-Time Airflow Model for Biopharmaceutical Manufacturing Facility Design

In biopharmaceutical manufacturing settings, ensuring the safety and purity of the final drug product is of utmost importance to patient safety. One critical aspect of manufacturing is environmental quality control by maintaining proper Heating, Ventilation, and Air Conditioning (HVAC) requirements in the facility as per the guidelines set by International Society of Pharmaceutical Engineers (ISPE), ISO 14644, and other US and EU guidelines together with establishing a common basis for compliance with current good manufacturing practices (cGMP).

Historically, venting and airflow has been characterized experimentally using smoke studies during clean room and facility qualification to identify any potential sources of contamination. Real time environmental monitoring tools are often deployed to monitor contaminations and assess specific threats to the manufacturing environment which can have a potential impact to product/ patient safety. These tools are effective but resource intensive to implement and expensive to fix if issues are found during system qualification. Additionally, the approach can still have uncertainty associated with airflow patterns and particulate levels in the facility design if traditional design tools are used.

In this work we use transient computational fluid dynamics (CFD) technology based on Lattice-Boltzmann (LB) method to accurately simulate the release and movement of particles in real time, allowing for a more efficient and cost-effective evaluation of airflow patterns. LB is a highly efficient and accurate method for simulating fluid flow and has become increasingly popular in recent years due to its ability to leverage GPU-based computing architectures. By utilizing GPUs, LB simulations can run orders-of-magnitude faster than traditional CFD approaches on CPUs, making it a highly efficient and cost-effective alternative. LB works by discretizing the domain into a lattice and simulating the movement of fluid particles as they collide with each other and the lattice boundaries. By tracking the movement of these particles, LB simulations can accurately model the complex airflow patterns.

By utilizing LB method and integrating it with facility design by using Building Information Modeling (BIM), we have created an airflow digital model of the facility that simulates the real-time air and particulate transport in a manufacturing facility. We are using this model to optimize HVAC design by optimizing the supply and exhaust register locations to minimize steady state particulate levels. Room velocities, temperature gradients, pressure gradients, particulate levels and particulate recovery times are predicted to ensure compliance with room classification criteria during dynamic (operating) conditions.

Overall, the transient CFD models developed in this work will provide an accurate science-based understanding of airflow requirements during the early stages of HVAC and facility design and reduce the overall costs/ timeline for implementation of CapEx projects. Integration of this technology in the critical areas of the facility will be used as risk-assessment tool to provide safe, compliant, and functional systems and facilities for biomanufacturing.