Research Projects
The main research of our lab is to perform high-performance computational simulations and to perform population-based lung imaging analysis. The ultimate aim is to develop the efficient intervention via imaging-based clustering, and to understand underlying structural and functional mechanism of lung diseases.
？
Collaborative Teams and Projects:
-
Ministry of Environment (환경부)
-
Korea Electric Power Corporation (KEPCO, 한국전력)
-
National Research Foundation of Korea via Ministry of Education (연구재단, 교육부)
-
Subpopulations and intermediate outcome measures in COPD study (SPIROMICS)
-
Chonbuk National University Hospital (CNUH)？
-
Kangwon National University Hospital (KWNUH)
-
Seoul National University of Science and Technology CFD Lab.
Complex Internal Flows, Viscous dissipation, Particle delivery with advanced CFD techniques
There are many engineering problems associated with internal flows, so-called pipe flows. While Re~2000 is known as a criterion of turbulent transition, actual flows in complex geometry such as bifurcation and curved shape are not clearly understood with existing knowledge of fluid mechanics.
With advanced CFD techniques along with automated mesh generation and a large eddy simulation model (LES), we solve not only biological problems, but also industrial problems. The MPI-parallel CPU-based computation using unstructured grids allows us to tackle highly complicated and large fluid systems.
</wix-image>
</wix-image>Image analysis, Registration and Cluster analysis using Quantitative Computed Tomography (QCT) images
QCT-based imaging metrics derived from post-processed image analysis and registration are objective tools to evaluate structural and functional alterations at global and local scales. However, these metrics may be used in an erroneous way due to inter-subject and inter-site variability. These issues could be resolved with novel normalization based on the perspective of mechanical engineering.
A set of QCT imaging variables can be used to perform a machine learning technique, deriving clinically meaningful disease groups. The ultimate aim of these grouping is to develop therapeutic intervention and to link with CFD characteristics based on respective clusters.
This project inherently pursues a large data associated with multi-center study. We collaborate SARP, SPIROMICS, SNUH and more to employ asthma and COPD data.
</wix-image>Pulmonary Air Flows Using Both 3D and 1D CFD Techniques
Human lungs with successive branching structures breathe about 15 times per 1 minute including inspiration and expiration. Due to the multiscale structure, as the number of generation increases, air-flow regime is ranged from turbulent to micro flows. In order to simulate pulmonary air flows, reliable turbulent model and physiologically consistent boundary conditions are necessary. Parallel finite element method-based CFD technique is able to simulate complex air flows, especially for disease lungs such as asthma and chronic obstructive pulmonary disease.
By collaborating medical doctors, we try to utilize the CFD technique in a clinical setting as a diagnostic tool. We also improve this CFD technique to optimize for the variety of cases associated with internal flows.
Two-phase Flows with FEM, OpenFOAM, and ANSYS
Existing CFD method to solve two-phase flows with surface tension has been suffered due to spurious currents created by the force imbalance between pressure gradient and surface tension force. In addition, the existing continuum surface force (CSF) model has a severe restriction of time step due to the explicit treatment. We developed a semi-implicit consistent CSF model in a FEM domain. This approach successfully resolved the issues of spurious currents and temporal stability.
The approach developed in 2-D domain is going to be extended to 3-D domain to solve more practical problems. By adding thermal model to the surface tension, the effect of phase change is also considered. We collaborate this topic with a CFD lab. in Seoul National University Science and Technology.