Intelligent Microfluidics for Plasma Separation: Integrating Computational Fluid Dynamics and Machine Learning for...

Area of Science:

• Biomedical Engineering
• Microfluidics
• Machine Learning

Background:

• Efficient blood plasma separation is crucial for point-of-care diagnostics, especially in resource-limited settings.
• Conventional centrifugation methods are slow, resource-intensive, and not suitable for portable applications.

Purpose of the Study:

• To develop an "Intelligent Microfluidics" platform integrating machine learning (ML) and computational fluid dynamics (CFD) for optimized plasma separation.
• To demonstrate the platform's potential for rapid, scalable, and portable diagnostics.

Main Methods:

• Utilized COMSOL Multiphysics to model a trifurcation microchannel for plasma separation.
• Employed eight supervised ML algorithms, including Artificial Neural Networks (ANN) and k-Nearest Neighbors (KNN), for performance prediction.
• Simulated fluid dynamics mimicking blood viscosity and density with optimized boundary conditions.

Main Results:

• Achieved high plasma yields of 90-95% across a range of inflow velocities.
• ANN demonstrated the highest predictive accuracy with R² = 0.97.
• The ML-enhanced microfluidic system showed superior performance and computational efficiency compared to traditional methods.

Conclusions:

• The intelligent microfluidic platform enables efficient, rapid plasma separation for real-time diagnostics.
• The scalable and portable design is ideal for healthcare in remote or resource-constrained areas.
• This approach lays the foundation for next-generation portable diagnostic technologies.