建筑环境非稳态问题的快速CFD模拟方法研究

Computational Fluid Dynamics (CFD) has been widely applied on research of building environment related flow and heat transfer problems. For unsteady state problem, commonly used CFD method lacks of validation of time-dependent result; moreover, since iteration within each time-step is employed to obtain simulation result, great computational cost and slow simulation speed is resulted. This research will explore non-iteration CFD algorithm to simulate typical building environment related unsteady state problems to improve computing simulation efficiency of CFD, as well as validate time-dependent solution to check the accuracy. The effect of related parameters, including velocity-pressure decoupling algorithm, turbulence model, grid scale and time-step size on computing speed and simulation accuracy of unsteady problems will be systematically tested. A fundamental theoretical CFD model will be obtained through comprehensive analysis. Numerical experiment and benchmark physical experiment on different flow mechanisms in typical building environment will be conducted to validate and improve the model. Validated model will be applied to unsteady state problem of real building environment cases to further verify the applicability, accuracy and speed. The outcome of this research will greatly improve CFD simulation speed on unsteady state problem of typical building environment, with satisfaction of basic accuracy requirement of such application. The research will also reveal the correlation between parameter input and simulation accuracy in unsteady state problem simulation of building environment.

计算流体力学（CFD）方法在建筑环境相关的流动和换热问题研究中有着广泛应用。对于非稳态问题的模拟，常用的CFD模型对时间相关解缺少必要的校核；在每个时间步长内，通过迭代计算获得模拟结果的方式，运算量大、速度慢。因此，本项目拟针对建筑环境中典型的非稳态问题，探索利用非迭代型CFD算法进行模拟，提高计算效率，并对时间相关的解进行校核，全面系统的验证压力-速度解耦算法、湍流模型、网格尺度和时间步长对于非稳态问题求解速度和精度的影响。课题将通过理论分析获得快速非稳态CFD算法的理论模型，并针对典型建筑环境中流动形成的不同机理，通过数值实验和基准物理实验的方式，对理论模型进行校核和改进，以获得快速非稳态CFD算法的校核模型。获得的模型还将通过建筑环境中实际的非稳态案例进行验证和必要改进。该研究将显著提高CFD方法在求解建筑环境非稳态问题上的运算效率，揭示非稳态问题求解精度和速度与模型参数的内在联系。

计算流体力学（CFD）方法在建筑环境相关的流动和换热问题研究中有着广泛的应用。对于非稳态问题的模拟，常用的CFD模型对时间相关的解缺少必要的校核；在每个时间步长内，通过迭代计算获得模拟结果的方式，运算量大、速度慢。本项目针对建筑环境中典型的非稳态问题，开发了一种非迭代型压力速度解耦算法-半拉格朗日PISO（SLPISO）算法进行模拟，提高模拟计算效率，并对时间相关的解进行校核，全面系统的验证压力-速度解耦算法、湍流模型、网格尺度和时间步长对于非稳态问题求解速度和精度的影响。该SLPISO算法在开源软件openFoam中执行，用于典型算例的验证和校核，改进了标量场（温度场和浓度场）的模拟计算效率，改善了算法的质量守恒特性。课题针对典型建筑环境中流动形成的不同机理，通过文献检索、数值实验和基准物理实验的方式，获得了校核所用的基准数据，对开发的算法通过建筑环境中实际的非稳态案例进行验证。共获得其它研究团队的3个算例和本课题的3个物理实验/数值实验的基准实验数据，在验证算法精度和速度的基础上，测试了时间步长设置规律，给出了大于库朗数限制的时间步长，验证了常数粘度法与常用涡粘模型在速度和精度方面的性能，给出了常见室内环境问题的网格数量建议值，为全面改进非稳态问题的CFD模拟效率提供了理论和实践方面的指南。