Some Research On High Performance Algorithm For Numerical Simulation Of Human Hemodynamic

Recently,the numerical simulation is widely used in engineering applications with the development of computer and numerical method.It has become the third method of scientific research besides the theoretical analysis and experimental research.The numerical simulations can replace some physical experiments which are costly and difficult to duplicate,and it would be better and faster to solve many scientific and engineering problems.Numerical simulation has been widely used in biomedical field,and it has become an important method in the human hemodynamics studies.The numerical simulation of human hemodynamics began in the 1980 s,and a lot of research results have been obtained during the last few decades.However,there are still some shortages in existing researches,such as the fidelity of vascular model of existing research is not high enough,there is still disparity to the real human blood vessels,and so on.Most of the existing research is based on the commercial CFD software which makes some parallel efficient algorithms are difficult to be used,and it would takes a long time when the scale of simulation problems is large.All in all,further research is needed.This thesis developed a set of high performance numerical simulation system for human hemodynamics simulation.The system includes the construction of 3D artery geometry from MRI images,unstructured mesh generation,the discretization and solution of the fluid flow equations,and the analysis of the results.We mainly focus on the flow equations solver part in this thesis and verify the reliability and efficiency of the system by solving a real patient case of stroke(including preoperative and postoperative cases)which is provided by Beijing Tiantan Hospital.The results show that the computed values,including the pressure,the velocity and the wall shear stress are consistent with the clinical conclusion.In addition,we also performed a numerical simulation of cardiovascular hemodynamics.Although there is a big gap between the simulation result and real data,it’s enough to show that the numerical simulation can be a good method in cardiovascular hemodynamics studies.Regarding to the parallel performance of the method,an almost linear speedup is obtained with up to 128 processor cores for a case with mesh elements62.12?10 and the parallel efficiency is still around 72.43%.