论文部分内容阅读
目的为降低梯形截面丝支架植入颅内囊状动脉瘤后瘤腔破裂的风险,对支架丝截面底边长度进行优化设计。方法构建38种不同底边长度的梯形截面支架及其植入动脉瘤后的有限元模型,利用流固耦合数值模拟计算得到38组最大瘤腔壁面压力梯度值,并应用广义回归神经网络及遗传算法对梯形截面底边长度进行优化,使植入支架后最大瘤腔壁面压力梯度降到最低。结果优化结果显示,相对传统矩形截面支架,优化后支架将最大瘤腔壁面压力梯度降低了7.86%。结论广义回归神经网络与遗传算法结合可以很好地解决支架优化方面的问题。
Objective To reduce the risk of tumor rupture after trabecular meshwork implantation in intracranial saccular aneurysms, and to optimize the length of the bottom of stent cross-section. Methods Thirty-eight trapezoidal cross-section scaffolds of different base lengths and their finite element models after implantation of aneurysms were constructed. The maximal wall pressure gradients in 38 groups were calculated by means of fluid-structure interaction numerical simulation. The generalized regression neural network and genetic algorithm The algorithm optimizes the length of the base of the trapezoid cross-section, so that the pressure gradient of the largest tumor cavity wall after implantation of the stent is minimized. Results The optimized results showed that the optimized stent reduced the pressure gradient of the largest tumor wall by 7.86% compared with the traditional rectangular cross-section stent. Conclusion The combination of generalized regression neural network and genetic algorithm can well solve the problem of stent optimization.