Development of procedures for reduced mechanism generation for prediction of particle formation in turbulent reacting flows

如果没有今天计算资源的进步，过去二十年来燃烧科学和技术的巨大进步是不可能的。我们现在能够越来越自信地模拟和模拟简化的燃烧现象，使用简单的燃料进行科学研究，而发动机制造商每天都在为他们的开发工作使用量身定制的燃烧代码。由于实际装置中的燃烧现象本质上是湍流的，因此预测燃烧的挑战是在现实环境中描述复杂的燃烧动力学。因此，在实现湍流燃烧预测可靠性的目标中，要求以精确和计算适应性的方式描述化学。此外，如果污染物浓度较低，模型预测的精度要求极高，就像在现代高度优化的设备(如精益、预混、预汽化（LPP）燃气轮机和新概念（如均质装药压缩点火（HCCI）发动机）的排气中发现的那样。预测模拟的新挑战与替代清洁燃料（如生物燃料）的未知行为有关。然而，替代燃料的性质往往非常复杂，例如生物质燃料。正在开发越来越多利用生物质的可再生能源项目，这些项目利用农业和工业的废物。虽然生物质燃料（碳中性燃料）的总体温室气体排放量通常较低，但需要解决与这种燃料有关的高浓度重金属和大量颗粒形成的问题。他们对可能妨碍这些设备高效、清洁和持久运行的沉积物和有害产品负责。至关重要的是，科学界致力于减少反应机制，迫切解决燃烧过程中颗粒形成的影响。当使用计算要求很高的模拟工具时，为了达到实际的计算时间，减少化学方案是必要的。事实上，即使是中等复杂的碳氢化合物（C7和C8）的详细机制也包含数百种通过数千种反应进行的反应。此外，碳氢化合物的氧化还具有形成高活性自由基的特点，其反应时间比主要自由基短得多。由此产生的时间尺度上的巨大差异导致控制化学演化的微分方程（ode）系统的刚度，这导致模拟只能按照最短的时间步长进行。因此，提出的工作解决了对紧凑但可理解的反应动力学模型的迫切需求，也能够预测粒子形成的前体。城市地区的有害微粒排放主要来自内燃机的燃烧。颗粒的形成位于非预混燃烧期间的富燃料区域，如柴油发动机，或在燃烧期间由于不均匀性而形成的小燃料富口袋，如汽油发动机，特别是在直喷模式下运行。由于大多数实际燃烧过程都涉及气体的湍流混合，因此湍流流场与化学过程之间的相互作用非常重要。所建立的简化模型将用于研究中等湍流混合、混合和化学不能解耦的条件下优先扩散的影响，以及柴油机燃烧末期形成大部分颗粒排放的典型条件。对湍流影响下优先扩散对粒子形成的影响的这一重要的新认识，反过来将为进一步的模型开发奠定基础，以实现实际发动机仿真的商业代码。

The tremendous progress in combustion science and technology in the past two decades would have been impossible without todays advances in computational resources. We are now able to model and simulate with growing confidence simplified combustion phenomena, using simple fuels for scientific studies, while engine manufacturers employ tailored combustion codes for their development work on a daily basis.Since combustion phenomena in practical devices are turbulent in nature, the challenge for predictive combustion is to describe complex combustion kinetics in a realistic environment. Hence, within the goal of achieving predictive reliability of turbulent combustion lies the requirement that the chemistry is described in an accurate and computationally adaptable manner. Also, the demanded accuracy of model predictions is extremely high if the concentrations of the pollutants are low, as found in the exhausts of modern, highly optimised devices, such as lean, premixed, pre-vaporised (LPP) gas turbines and in new concepts like the homogenous charge compression ignition (HCCI) engine. New challenges for predictive simulations are related to the yet unknown behaviour of alternative clean fuels, such as bio-fuels. However, alternative fuels are often very complex in nature, such as biomass fuels. An increasing number of renewable energy projects using biomass are under development, using waste products from agriculture and industry. Although overall green-house gas emissions are typically low for biomass fuels (carbon neutral fuels), concern need to be addressed regarding high concentrations of heavy metals and significant particle formation associated with such fuels. They are responsible for deposits and harmful products that may hinder the efficient, clean and durable run of these devices. It is crucial that the scientific community working with reduced reaction mechanisms urgently addresses the impact of particulate formation in combustion processes. When using computationally demanding simulation tools, reduced chemical schemes are necessary in order to achieve practical computing times. Indeed, detailed mechanisms for even moderately complex hydrocarbons (C7 and C8) contain hundreds of species reacting through thousands of reactions. Furthermore, hydrocarbon oxidation is also characterized by the formation of highly reactive radicals reacting on much smaller time scales than the major species. The resulting dramatic difference in time scales consequently result in stiffness in the system of differential equation (ODEs) governing the chemical evolution, which causes the simulations to progress only according to the shortest time steps. Hence, the proposed work addresses the urgent need for compact, yet comprehensible reaction kinetic models, also capable of predicting the precursors for particle formation.Harmful particulate emissions in urban areas largely originate from combustion in IC engines. The formation of the particles is located in fuel rich regions during non-premixed combustion such as in diesel engines, or in small fuel rich pockets due to inhomogeneities during combustion e.g. in petrol engines, in particular run in direct injection mode. As most practical combustion processes involve the turbulent mixing of gases, the interactions between the turbulent flow field and the chemical processes are important. The formulated reduced model will in turn be used to study the effect of preferential diffusion under moderate turbulent mixing, conditions under which mixing and chemistry can not be decoupled, and typical conditions at the end of the combustion phase of in diesel engines when most of particulate emissions are formed. This important new understanding of the effect of preferential diffusion on particle formation under the influence of turbulence will in turn represent the basis of further model development towards implementation into commercial codes for realistic engine simulation.