Resolving discrepancies in detailed kinetic models of combustion via automated transition state theory calculations

1605568威斯汀为了开发更清洁和更高效的发动机，并更好地利用石油衍生燃料和替代燃料，工程师们需要更好地了解燃烧。能够准确预测不同燃料在不同条件下燃烧情况的计算机模型将帮助我们开发新发动机和新燃料。这项研究的目标是预测、详细的燃烧动力学模型，在该模型中，数千个相关的化学反应得到准确计算。作为迈向这一目标的一步，这项研究将检测并纠正目前在详细动力学模型中使用的反应速率中的主要差异，并在这样做的过程中创建一个反应过渡态数据库，以及预测它们的算法。拟议的工作将使用几种新的技术来识别和纠正目前隐藏在详细燃烧动力学模型中的错误、不确定性和近似值。该项目将消除人类在进行化学动力学量子力学计算时的瓶颈，使高性能计算在未来能够有效地用于准确计算反应速率表达式。高通量的反应动力学计算已被确定为“21世纪运输燃料清洁和高效燃烧的基础研究需要”，并被燃烧能源前沿研究中心确定为“重要的重大挑战”。这个为期三年的项目提出：(1)在开源的动力学模型建立软件Reaction Machine Generator(RMG)中自动执行基于量子力学(QM)的过渡态理论(TST)计算与燃烧相关的反应；(2)使用新开发的动力学模型导入工具来识别最近发布的燃烧模型中的每个基元反应；(3)使用第一步中的自动化方法计算第二步中的反应速率，创建一个公共数据库，其中包括反应、文献中的速率、由TST计算的速率、QM计算结果和过渡态几何形状；以及(4)识别已发表的模型中的反应速率与通过TST计算的反应速率之间的差异，并量化这些差异对模型预测的影响。此外，我们还将开发一套适用于本科化学工程课程的基于Python的相关教材，将燃烧科学介绍给更广泛的受众。一名博士生和几名本科生将在这个项目中获得宝贵的研究经验和培训。

1605568 WestIn order to develop cleaner and more efficient engines, and to make better use of both petroleum-derived and alternative fuels, engineers need a better understanding of combustion. Computer models that can accurately predict how different fuels burn in different conditions will help us develop new engines and new fuels. The goal of this research is predictive, detailed kinetic modeling of combustion, in which many thousands of relevant chemical reactions are calculated accurately. As a step towards this goal, the research will detect and correct major discrepancies in the reaction rates currently used in detailed kinetic models, and in so doing create a database of reaction transition states, as well as algorithms to predict them. The proposed work will use several novel techniques to identify and correct mistakes, uncertainties, and approximations currently hidden throughout detailed kinetic models of combustion. The project will remove the human bottle-neck in performing quantum mechanical calculations of chemical kinetics, enabling effective use of High Performance Computing for accurate calculation of reaction rate expressions in the future. This high-throughput calculation of reaction kinetics has been identified as a "basic research need for clean and efficient combustion of 21st century transportation fuels", and by the Combustion Energy Frontier Research Center as an "important grand challenge".This three year project proposes to: (1) automate the performance of quantum mechanics (QM) based Transition State Theory (TST) calculations for combustion-relevant reactions in the open-source, kinetic model building software Reaction Mechanism Generator (RMG); (2) use a newly developed kinetic model importer tool to identify every elementary reaction published in recent combustion models; (3) use the automated methods from step one to calculate the rates of reactions from step two, creating a public database of reactions, rates from the literature, rates calculated by TST, QM calculation results, and transition state geometries; and (4) identify discrepancies between reaction rates in published models and those calculated via TST, and quantify the effect these discrepancies have on the model predictions. Additionally, we will develop a related suite of Python-based teaching materials suitable for undergraduate chemical engineering curriculum, to introduce combustion science to a wider audience. A PhD student and several undergraduate students will gain valuable research experience and training whilst working on this project.