论文标题

部分可观测时空混沌系统的无模型预测

Utilizing bifurcations to separate particles in spiral inertial microfluidics

论文作者

Valani, Rahil, Harding, Brendan, Stokes, Yvonne

论文摘要

储层计算是预测湍流的有力工具,其简单的架构具有处理大型系统的计算效率。然而,其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间,然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用,该系统模拟了湍流的若干特征。我们表明,使用径向基函数作为非线性投影器,即使只有部分观测并且不知道控制方程,也能稳健地捕捉复杂的系统非线性。最后,我们表明,当测量稀疏、不完整且带有噪声,甚至控制方程变得不准确时,我们的网络仍然可以产生相当准确的预测,从而为实际湍流系统的无模型预测铺平了道路。

Particles suspended in fluid flow through a closed duct can focus to specific stable locations in the duct cross-section due to hydrodynamic forces arising from the inertia of the disturbed fluid. Such particle focusing is exploited in biomedical and industrial technologies to separate particles by size. In curved ducts, the particle focusing is a result of balance between two dominant forces on the particle: (i) inertial lift arising from small inertia of the fluid, and (ii) drag arising from cross-sectional vortices induced by the centrifugal force on the fluid. Bifurcations of particle equilibria take place as the bend radius of the curved duct varies. By using the mathematical model of Harding, Stokes, and Bertozzi [1], we illustrate via numerical simulations that these bifurcations can be leveraged in a spiral duct to achieve large separation between different sized particles by transiently focusing smaller particles near saddle-points. We demonstrate this by separating similar-sized particles, as well as particles that have a large difference in size, using spiral ducts with square cross-section. The formalism of using bifurcations to manipulate particle focusing can be applied more broadly to different geometries in inertial microfluidics which may open new avenues in particle separation techniques.

扫码加入交流群

加入微信交流群

微信交流群二维码

扫码加入学术交流群,获取更多资源