Computationally efficient LES-TPDF turbulent combustion models

计算高效的 LES-TPDF 湍流燃烧模型

• 批准号: 319178666
• 负责人: Dr. Yipeng Ge, Ph.D.
• 金额: --
• 依托单位: Institut für Thermodynamik (LRT-10)
• 依托单位国家: 德国
• 项目类别: Research Grants
• 财政年份: 2016
• 资助国家: 德国
• 起止时间: 2015-12-31 至 2020-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/319178666?language=en
• 关键词: Computationally; efficient; LES; TPDF; turbulent

TThe combustion in many technical burners occurs in turbulent flow. Due to the overlap of chemical and turbulent and time scales and because of the non-linear dependence of the reaction rate on the concentrations of chemical species, the turbulence chemistry interaction must be taken into account e.g. by using transported transported probability density functions (TPDF's). The turbulent flow can be calculated very well using Large Eddy Simulations (LES), however usually not all spatial scales of chemical reactions can be resolved. The direct coupling of TPDF's with LES is computationally intensive because the TPDF must be represented by many stochastic particles in each LES-computational cell. The calculation of the evolution of the chemical species is the computationally most expensive part of the code. The aim of this project is to develop methods that provides a similar quality of prediction of the evolution of chemical species at a significantly reduced computational effort. Two related TPDF approaches will be developed, which take into account the close correlation between the mixture fraction and species. Basis in both cases is a LES for the simulation of turbulent flow field. In one approach the TPDF are represented with a low number of stochastic particles (sparse Lagrangian method). The mixture model uses larger areas in physical space and mixes preferably particles of similar mixture fraction. In the second approach the TPDF is derived from a RANS simulation, which is performed using the time-averaged LES-flow field. The RANS simulation is cheaper since it needs only be stationary and a coarser computational grid can be applied. In both cases the feedback to the LES occurs by means of a density conditioned on the mixture fraction. To further reduce the computational time, the solution of chemical evolution on GPU's and the use of tables with systematically reduced chemical models are examined.

在许多技术燃烧器中，燃烧发生在湍流中。由于化学和湍流和时间尺度的重叠，并且由于反应速率对化学物质浓度的非线性依赖性，必须考虑湍流化学相互作用，例如通过使用传输的传输概率密度函数（TPDF）。大涡模拟（LES）可以很好地计算湍流，但通常不是所有的空间尺度的化学反应可以解决。TPDF与LES的直接耦合是计算密集型的，因为TPDF必须由每个LES计算单元中的许多随机粒子表示。化学物种演化的计算是代码中计算成本最高的部分。该项目的目的是开发方法，提供了一个类似的质量预测的化学物种的演变，在一个显着减少计算工作。将开发两种相关的TPDF方法，考虑到混合物分数和物种之间的密切相关性。在这两种情况下的基础是一个大涡模拟湍流场。在一种方法中，TPDF用少量随机粒子表示（稀疏拉格朗日方法）。混合物模型在物理空间中使用较大的区域，并且优选地混合相似混合物分数的颗粒。在第二种方法中，TPDF来自RANS模拟，这是使用时间平均LES流场进行。RANS模拟更便宜，因为它只需要是静止的，并且可以应用更粗糙的计算网格。在这两种情况下，反馈到LES发生的密度条件下的混合物分数的装置。为了进一步减少计算时间，在GPU上的化学演化的解决方案和使用表与系统地减少化学模型进行检查。