Efficient multi-fidelity computation of blood coagulation under flow

Area of Science:

• Biophysics
• Computational Biology
• Biochemistry

Background:

• Blood clot formation is vital for hemostasis but imbalanced coagulation can cause severe disorders.
• The coagulation cascade, regulating thrombin and fibrin formation, is modeled using complex partial differential equations (PDEs).
• Simulating these large, multi-scale PDE systems computationally is challenging due to their complexity.

Purpose of the Study:

• To develop a more efficient computational strategy for simulating the coagulation cascade.
• To reduce the computational cost of modeling blood clot formation in complex flow scenarios.
• To enable advanced analyses of coagulation dynamics in intricate biological systems.

Main Methods:

• Transformed governing PDEs into ordinary differential equations (ODEs) based on blood residence time.
• Employed Taylor expansion around the zero-diffusivity limit to derive species concentrations from residence time statistical moments.
• Developed new PDEs to govern these statistical moments, creating a multi-fidelity model.

Main Results:

• The multi-fidelity strategy replaces N PDEs with N ODEs and p PDEs for statistical moments, offering a speedup of over N/p.
• Computational cost becomes independent of the number of chemical species in large meshes.
• Low-order models (p=1, p=2) demonstrated favorable accuracy, with thrombin concentration deviating by <20% (p=1) and <2% (p=2) after 20 cycles.

Conclusions:

• The proposed multi-fidelity approach significantly enhances the efficiency of coagulation cascade simulations.
• This method allows for accurate modeling of coagulation in complex geometries and flow conditions.
• The strategy is generalizable to other flow-affected reacting systems, advancing scientific understanding.