I-Corps: High-fidelity Simulation Software for Microfluidics

微流体系统越来越多地用于几种人类疾病的诊断和治疗。例如，它们在各种癌症和艾滋病毒的低成本、即时诊断方面具有巨大的潜力。这种系统成功的关键是精确控制和操纵单个血细胞的能力。虽然到目前为止，它们的设计主要基于化学和详尽的实验室测试，但新的预测计算机模拟工具正在出现。目前的商业软件包提供了优秀的模拟工具来分析简单（牛顿）流体通过任何给定的微流体装置的流动。然而，在涉及诸如细胞、气泡或胶囊的可变形颗粒的流动的情况下，其范围极其有限。基于数值算法的最新进展，该项目将开发能够模拟真实血细胞浓度的新型模拟技术，从而实现微流体设备的快速原型制作。因此，该项目有可能推动低成本诊断工具和患者特定治疗策略的进一步创新。颗粒流模拟的主要困难在于流体动力学相互作用是长程的，从而导致高计算费用。此外，时间步进计划遭受损失的准确性和稳定性，由于几乎接触的表面随时间变化的颗粒流动。这个I-Corps团队开发的求解器已经能够（i）通过使用快速N体算法在线性时间内计算细胞和通道之间的相互作用，（ii）通过使用频谱精确的数值方法提供高保真度的结果，以及（iii）通过使用克服刚度的时间推进方案进行长时间模拟。该项目的主要目标是建立一个模拟套件，适用于任何给定的芯片几何形状，边界条件和细胞浓度。

Microfluidic systems are increasingly being used in diagnostics and therapeutics of several human diseases. They hold tremendous promise for low-cost, point-of-care diagnosis of various types of cancer and HIV, for instance. Key to the success of such systems is the ability to precisely control and manipulate individual blood cells. While their design so far is largely based on heuristics and exhaustive lab testing, new predictive computer simulation tools are emerging. Current commercial software packages provide excellent simulation tools to analyze flow of simple (Newtonian) fluids through any given microfluidic device. However, in the case of flows involving deformable particles such as cells, bubbles or capsules, their scope is extremely limited. Based on recent advances in numerical algorithms, this project will develop novel simulation technology capable of simulating realistic blood cell concentrations, enabling rapid prototyping of microfluidic devices. Thereby, this project has the potential to drive further innovations in low-cost diagnostic tools and patient-specific therapeutic strategies. The main difficulty with particulate flow simulations lies in the fact that the hydrodynamic interactions are long-range, thereby, leading to high computational expense. Furthermore, the time-stepping schemes suffer from loss of accuracy and stability due to nearly touching surfaces that vary in time as the particles flow. The solvers developed by this I-Corps team are already capable of (i) computing the interactions between cells and channels in linear time via the use of fast N-body algorithms, (ii) delivering results with high-fidelity via the use of spectrally-accurate numerical methods and (iii) performing long-time simulations via the use of stiffness-overcoming time-marching schemes. The primary goal of this project is to build a simulation suite that works for any given chip geometry, boundary conditions and cellular concentrations.