Visualization: Assimilation of experimental and computational fluid dynamics
Traditionally, imaging has focused on (time-resolved) acquisition of morphological information. Using MRI, blood flow velocities can be measured in the cardiovascular system. Restrictions in measurement times have limited these scans to 2D, but recent advances in software and hardware allow for time-resolved 3D approaches (4D flow MRI). Analysis of this wealth of data is currently time-consuming and user-dependent.
This projects aims to facilitate and quantify visualization of cardiovascular blood flow. The project can be divided into 2 sub projects, which will be performed concurrently:
a) Assimilation of experimental and computational fluid dynamics.
Today, time-varying 3D velocity fields in the cardiovascular system are either measured with 4D flow MRI or simulated with computational fluid dynamics (CFD). The main advantage of measurements is of course that they reflect the actual status of the subject being studies. The main advantage of simulations is that they provide superior spatial and temporal resolution, and are capable of resolving the smallest scales of the flow. We aim to integrate these techniques. By generalizing the CFD approach, we seek to obtain a method that permits integration of sparse measured velocity data in the CFD simulations. This may allow for enhancement of measured velocity data by filling up gaps, increasing resolution and decreasing noise. With the extended integration of measured data in the simulations this approach may also lead to more accurate CFD results.
b) Visualization of cardiovascular blood flow.
Here, clinically important flow features of 4D flow MRI data of the whole heart and large vessels will be visualized. The resolution of the MRI data might be improved by integration of CFD simulations as described above. Abnormal flow (e.g. duct flow between heart chambers, abnormal flow directions through valves, elevated turbulence intensity) will be emphasized in the visualizations. Laminar and turbulent flow patterns and structures will be visualized and quantified by e.g. Lagrangian Coherent Structures (LCS) or Proper Orthogonal Decomposition (POD).
The project is financed by Swedish e-Science Research Centre (SeRC) Visualization, as a strategic research area (SRA) funded by Vetenskapsrådet and CENIIT.