CAREER: Development of a Hierarchy of Optimal LES Models for Turbulent Spray Combustion

湍流喷雾燃烧是一种常见的现象，出现在许多实际设备中，如飞机发动机、燃气轮机、化学反应器等。喷雾燃烧的预测模型对于估计这些系统的安全有效使用的操作包线至关重要。然而，考虑到喷雾燃烧过程的多尺度多物理场特性，开发此类模型是一项巨大的挑战。近年来，大涡模拟（LES）已成为替代传统的雷诺平均纳维-斯托克斯（RANS）燃烧模拟方法的一种可行方法。在LES中，大尺度运动直接计算，而小尺度过程则建模。虽然LES在预测远离熄灭的单相火焰方面显示出相当大的准确性，但模拟更复杂的过程（如喷雾燃烧或火焰熄灭）面临着与基于ranss的方法相同的难以克服的困难。造成这种僵局的主要原因是，LES燃烧模型几乎是相应的RANS模型的不变版本。这项工作的目的是开发新一代的燃烧模型，专门为LES的目标是准确预测所有的空间统计的流动配置。新的框架，称为条件过滤概率密度函数（FPDF）方法，将为理解燃烧建模中的建模和数值误差提供统计一致的方法。PI将与悉尼大学的实验人员和陶氏化学公司的研究人员密切合作，严格测试这里开发的模型。为了确保及时传播研究进展，计划了一个多层次的教育计划。将开发一个基于网络的学习平台，提供对模型、计算数据库和研究代码的访问。课堂学习的互动工具也将在此框架下开发。这些工具将用于培训工业研究人员和学生在计算燃烧科学的最新进展。此外，还将通过该项目支持高中生实习计划。积极的学生，特别是来自代表性不足的群体的学生，将有机会参与研究活动并与德州大学的学生互动。

CBET-0747427, RamanTurbulent spray combustion is a common phenomenon occurring in a number of practical devices such as aircraft engines, gas turbines, and chemical reactors. A predictive model for spray combustion is vital in estimating an operational envelope for the safe and efficient use of these systems. Development of such models, however, is a tremendous challenge given the multiscale multiphysics nature of the spray combustion process. In the recent past, large-eddy simulation (LES) has emerged as a viable alternative to conventional Reynolds-Averaged Navier-Stokes (RANS) approach to combustion modeling. In LES, the large-scale motions are directly computed while the small-scale processes are modeled. While LES has shown considerable accuracy in predicting single-phase flames far from extinction, modeling more complex processes such as spray combustion or flame extinction faces the same insurmountable difficulties as RANS-based approaches. The main reason for this impasse is that LES combustion models are nearly unaltered versions of the corresponding RANS models. The objective of this work is to develop a new generation of combustion models specifically for LES with the goal of predicting all spatial statistics of the flow configuration accurately. The new framework, termed conditionally-filtered probability density function (FPDF) approach, will provide a statistically consistent approach for understanding modeling and numerical errors in combustion modeling. The PI will closely interact with experimentalists from the University of Sydney and researchers from Dow Chemical Company to rigorously test the models developed here. To ensure timely dissemination of research advances, a multi-level educational program is planned. A web-based learning platform that provides access to models, computational databases, and research codes will be developed. Interactive tools for classroom learning will also be developed under this framework. These tools will be used to train both industrial researchers and students in the latest advances in computational combustion science. Further, an internship program for high-school students will be supported through this project. Motivated students, especially from underrepresented groups, will have an opportunity to participate in the research activities and interact with UT students.