基于Gillespie方法的血栓生成模拟方法

Thrombus Clotting Simulation Method Based on the Gillespie Method

血栓的生成是血液生理特性的体现,在虚拟手术和人体生理仿真中有着重要的作用.为了解决现有血液模拟方法不能体现血栓生成的现象以及模拟速度过慢的问题,提出一种基于Gillespie方法的血栓生成模拟方法.首先使用血栓生成模型对血液粒子建模,并且针对血液成分设定生物反应速率;然后基于Gillespie方法在核范围内进行生物反应的计算,由蒙特卡洛方法确定反应以及反应时间,将流体粒子转化为血栓粒子;最后将含有纤维蛋白的粒子位置固定,形成血栓.为了使模拟结果能够实时呈现,计算中使用了CUDA并行计算来加速模拟.分别对血栓形成模拟、血栓生成速度和程序运行效率进行分析,实验结果表明,文中方法实现了血液中血栓形成的快速模拟,避免了血栓生成的偏微分方程模型带来的高计算量,能够更好地应用在血液中血栓生成的实时模拟中;同时模拟血栓的生成对血液的流动路径造成影响,体现了血栓与血液的耦合作用;最后展现了几个逼真的血栓模拟的渲染结果.

Clotting of thrombus is a reflection of blood physiological properties and it plays an important role in virtual surgery and physiological simulation of a human body.In this paper,we propose a thrombus simulation method based on the Gillespie method.Firstly,the blood is modeled in view of thrombogenesis and physiology reaction rate are set according to reaction components.Then we determine the reactions and relevant time durations among particles based on the Gillespie method’s Monte Carlo steps within the kernel boundary before fluid particles are transformed to thrombus particles.At last,we fix the position of fibrous particles generated from the previous step and thrombus is clotted.In order to show the result in real time,CUDA parallel computing toolkit was utilized to speed up the computation.We analyzed thrombus clotting,clotting speed,and runtime efficiency through experiments.The results show that our method can simulate the clotting of thrombus efficiently,which avoids the high computational price of the Euler method.Meanwhile,the simulated thrombus would restrict the blood flow path,which shows us the coupling effect between blood and thrombus.We also rendered some realistic thrombus clotting pictures.