Development and application of lattice Boltzmann models for predictive large eddy simulation of turbomachinery aerodynamics

近年来，格子Boltzmann方法(LBM)已经成为风能工程和空气动力学的一种有趣的替代CFD方法。对于旋转机械，我们目前的实现已经比直接逼近不可压缩或弱可压缩的Navier-Stokes方程的传统求解器快得多。然而，在强湍流的预测模拟和边界层的模拟方面仍然存在挑战。本项目将集中于与单松弛时间格子Boltzmann方法和多松弛时间格子Boltzmann方法兼容的预测大涡模拟模型的开发和验证。这包括实现兼容的墙函数模型，这对于LBM固有的笛卡尔嵌入边界方法至关重要。除了标准基准外，还计划对水平轴风力涡轮机运行时的空气动力学进行大规模调查。所有开发都将在我们成熟的C++框架AMROC中进行，AMROC已经提供了全功能、动态自适应和完全并行的LBM实现，并对叶轮机械模型进行了动态流固耦合。

In recent years, the lattice Boltzmann method (LBM) has emerged as an interesting alternative CFD approach for wind engineering and aerodynamics. For rotating machinery, our current implementation is already significantly faster than conventional solvers that approximate the incompressible or weakly compressible Navier-Stokes equations directly. However, challenges remain in the predictive simulation of highly turbulent flows and the modeling of boundary layers.This project will concentrate on the development and validation of predictive large eddy simulation models that are compatible with single- and multi-relaxation time lattice Boltzmann methods. This includes implementation of compatible wall function models, which are of crucial importance for the Cartesian embedded boundary approach intrinsic to LBM. Beside standard benchmarks, large-scale investigations into the aerodynamics of operating horizontal axis wind turbines are planned. All developments will be carried in our mature C++ framework AMROC, which already provides a fully functional dynamically adaptive and fully parallel LBM implementation with on-the-fly fluid-structure coupling to turbomachinery models.