Dynamic Mode Decomposition Analysis of High-Fidelity CFD Simulations of the Sinus Ventilation

Abstract

Physiologically, sinus ventilation is a critical aspect of good functionality for human respiratory function, the understanding of which is still unclear. In this study we develop a method to measure sinus ventilation. Spatial and temporal features of the sinus recirculation are provided through dynamic mode decomposition (DMD). We associate the recirculation feature on the sinus epithelial surface through the wall shear-stress to the 3D airflow features corresponding to the sinus. Turbulent airflow simulations using a Large Eddy Simulation model were conducted in a patient who receives an endoscopic sinus surgery of the ethmoid + sphenoid sinuses. We analyze the flow rate through the ostium and the average wall shear-stress on the sinus epithelial surface. Based on the results, we then use a dynamic mode decomposition method on the sinus, which accurately measures the recirculation in the cavity. Drug delivery in the paranasal sinuses is challenging; therefore it is essential to understand well the airflow in this region. We suggest that DMD is a powerful method to analyze and understand this complex pathophysiology problem. This method has to be considered for the rhinologists as a milestone.