DNS-driven development of predictive LES models for gas turbine emissions

减少污染物排放是燃气轮机燃烧器发展面临的主要挑战之一。在这种情况下，大涡模拟（LES）有助于降低开发成本，加快开发进程。因此，本项目的目的是改进航空发动机和固定式燃气轮机的LES排放模型。目前的提案是一个2+2年项目的第二阶段。对于航空发动机，重点是烟灰的建模。在一期研究了烟尘的形成和氧化后，二期将进一步研究烟尘从富油区向贫油区的突破。当由于航空发动机的高湍流水平，湍流涡流将烟灰袋从燃料丰富的区域通过火焰输送到燃料贫乏的区域，以至于烟灰颗粒无法完全氧化时，就会发生烟灰突破。目前的LES模型不能以足够的精度描述这些事件，因此严重低估了煤烟排放。在本项目中，将扩展文献中现有的和在前一个项目阶段开发的最先进的LES模型，以考虑烟灰突破现象。这将分三步完成：1)为三个不同级别的煤烟突破生成DNS数据库；2)基于DNS的现有模型评价及考虑烟尘突破的模型扩展；3)通过LES对实验室规模的射流火焰配置和全模型燃烧器进行模型验证。在固定式燃气轮机中，一氧化碳（CO）排放将被调查。本阶段的目的是研究氢气（H2）加入甲烷（CH4）对CO排放的影响及其建模。首先是H2作为未来能源载体的重要性。在这里，我们将考虑在固定燃气轮机中向天然气中添加少量H2作为燃料。先前的研究表明，由于火焰速度的增加，即使是少量的H2也能显著减少CO的排放。对CH4/H2混合物的湍流燃烧进行了一系列的DNS实验。将对DNS数据进行分析，以扩展现有的CO模型，充分考虑H2添加的影响。完整的模型将在两个模型燃烧室的LES中使用，并用实验数据进行验证。这一项目的资金是由联邦政府和联邦志愿机构共同要求提供的。虽然两个拟议的工作包涉及不同环境中的污染物形成，但通过将工作包嵌入FVV的工业合作伙伴提供的项目支持中，可以建立牢固的联系。类似的建模方法和研究方法将用于CO和烟灰的预测，以便联合模型的开发可以导致污染物形成的通用描述。

The reduction of pollutant emissions is one of the main challenges in the development of gas turbine combustors. In this context, large-eddy simulations (LES), can contribute to reducing development cost and accelerating development processes. Therefore, the aim of this project is the improvement of emission models for LES of both aeroengines and stationary gas turbines. The present proposal is for the second phase of a 2+2-year project. For aeroengines, the focus is on modeling of soot. After soot formation and oxidation have been investigated in the first project phase, a better understanding for the breakthrough of soot from rich to fuel-lean regions is intended in the second phase. Soot breakthrough occurs when, due to the high turbulence levels in aeroengines, turbulent eddies transport soot pockets from fuel rich regions through the flame to fuel lean regions fast enough that a complete oxidation of soot particles cannot take place. Current LES models cannot describe these events with sufficient accuracy, thus significantly underpredicting soot emissions. In this project, state of the art LES models available in the literature and developed in previous project phases will be extended to consider soot breakthrough phenomena. This will be done in a three-step process: 1) generation of a DNS data base for three different levels of soot breakthrough; 2) assessment of current models based on the DNS and development of model extensions considering soot breakthrough; 3) model validation via LES of a lab-scale jet flame configuration and of a full model combustor. In stationary gas turbines, carbon monoxide (CO) emissions will be investigated. The aim in this phase is to study the effect of the hydrogen (H2) addition to methane (CH4) on CO emissions and their modeling. Starting point is the importance of H2 as future energy carrier. Here, we will consider the addition of small amounts of H2 to natural gas as fuel in stationary gas turbines. Previous work has shown the ability of even small amounts of H2 to significantly reduce CO emissions as a result of the increased flame speeds. A series of DNS of turbulent combustion of CH4/H2 mixtures will be performed. The DNS data will be analyzed to extend the present CO model to adequately consider the effects of H2 addition. The full model will be used in the LES of two model combustion chambers and validated with experimental data. Funding for this project is requested jointly from the DFG and the FVV. While the two proposed work packages address pollutant formation in different environments, strong connections are given by the embedding of the work packages in the project support provided by industrial partners of the FVV. Similar modeling approaches and research methodologies will be used for the prediction of CO and soot, so that the joint model development can lead to a generic description of pollutant formation.