Till innehåll på sidan

Mbah Chrameh Fru: Development of a pipeline for patient specific finite element model of the left ventricle of the human heart geared towards blood flow simulation

Tid: Ti 2014-10-21 kl 10.00

Plats: Room 3418, Lindstedtsvägen 25, 4th floor, Department of mathematics, KTH

Opponent: Erik Boström

Handledare: Johan Jansson

Exportera till kalender

It has been observed that in order to use a heart solver code developed at the computational technology laboratory at KTH with a data set provided by PHILIPS Research consisting of surface meshes of a whole heart, the surface mesh has to be converted to a volume mesh. The conversion is manual and time consuming. Therefore the purpose of this thesis is to develop algorithms and software tools such that it can be used in the automation (generating) of a finite element volume mesh of a left ventricle of a human heart based on the available PHILIPS Research data set. The developed model can later be used for simulating blood flowby solving the Navier-Stokes equation. The method used in generating the model is by deforming an a priori finite element volume mesh model to fit the extracted inner wall of the left ventricle, from the aforementioned data set. The deformation is done by solving a time dependent non linear elastic partial differential equation (pde) using finite element method. When solving this partial differential equation, the quality of each cell in the mesh is computed and for cells with low quality an extra stiffness is assigned to them making the vertices of the cells hard to move. This project starts with the characterization of the external field, and uses this external field as a component of the total force responsible for deforming the object. The method is then validated by deforming a two dimensional test case, for example a circle, into an ellipse and thereafter, deforming a three dimensional case, for example a sphere, into an ellipsoid. For each test case, two implementations of the pde was tested and found that for the two dimensional case, the deformation was better when the pde assigns extra stiffness to cells with low quality. For the three dimensional case the deformation was better when no extra stiffness was assigned. The method was then applied to the above mentioned data set, yielding results which are open for further improvements.