Computational methods for fluid flow, heat transfer, and microfluidics

该计划的目标是为流体流动和传热开发强大的计算方法，这些方法准确，经济，并跨越整个流动状态范围，并将这些方法应用于包括微流体“芯片实验室”在内的应用程序的设计周期。当精度达到一定程度时，至少对于中等马赫数，经济性、整个航程和设计应用问题都需要注意。与迄今为止使用的直接稀疏求解器相比，迭代、共轭梯度、多重网格和多块方法将被检查以应用于全速算法。为了进一步将马赫数范围扩展到高超音速范围，将需要更先进的技术。这些将包括考虑不同的线性化密度和他们的影响，收敛和稳定性。

The objectives of this program are to develop robust computational methods for fluid flow and heat transfer that are accurate, economical, and span the entire range of flow regimes, and to apply these methods in the design cycle for applications including the microfluidic 'lab-on-a-chip'. While the accuracy is in place, at least for moderate Mach numbers, the economy, entire range, and design application issues require attention. In contrast to direct sparse solvers used to date, iterative, conjugate gradient, multigrid and multiblock methods will be examined for application to the all-speed algorithm. To further extend the Mach range into the hypersonic range will require more advanced techniques. These will include the consideration of different linearizations of density and their influence on convergence and stability.