基于实时模拟信息反馈的湍流扩散火焰数值模拟研究

发布时间：2018-05-27 04:30

本文选题：火灾 + 实时模拟信息反馈　；参考：《中国科学技术大学》2016年博士论文

【摘要】：本质上来说,火灾是一种湍流燃烧现象。湍流燃烧的核心问题是研究湍流混合与复杂化学在多个时间-空间尺度上的相互作用关系。研究湍流混合与复杂化学的相互作用机理,有助于对火灾动力学更深层次的理解,有助于特殊工况下火灾模型的开发。研究湍流混合与复杂化学相互作用关系,首先需要海量的、包含耗散尺度湍流信息的反应流数据。本文采用一维湍流模型(ODT)来获取高保真数据集。ODT模型能够分辨Kolmogorov尺度上的湍流脉动信息,在耦合复杂化学计算时比多维直接模拟的计算代价要低得多。由于化学反应流是典型的变密度流,本文在原始ODT程序的基础上,综合考虑多维速度分量及密度变化对随机涡事件的影响,开发了变密度版ODT程序,并通过与实验值对比验证了变密度ODT程序的有效性。可分辨耗散尺度湍流信息的高精度ODT数值平台,为进一步分析湍流-化学相互作用关系和构建经验低维流形提供了健壮的数据库。研究湍流混合与复杂化学相互作用关系,还需要高效的、精准的反应流数据分析方法。本文构建了基于扩展定义的Damkohler数和拉格朗日示踪方法的高精度计算燃烧分析平台,从时间尺度关系出发,解耦湍流混合与化学反应的相互作用关系。本文应用ODT程序数值模拟了CO合成气湍流扩散燃烧过程,结合化学爆炸模式分析(CEMA)理论,建立了基于Damkohler数的湍流扩散火焰局部熄火重燃情节分析方法。结果显示,熄火区可以由大的、负的Damkohler数表示。此外,本文还从时间演化角度对拉格朗日示踪单元的爆炸指数(EI)进行了深入讨论。通过对熄火持续时间的区分,定义了两种不同的湍流扩散火焰局部熄火情节。通过对重燃过程爆炸指数的分析,定义了两种不同的湍流扩散火焰重燃机理：预混火焰传播重燃情节对应着较长时间的预热过程及较高的温度爆炸指数,且重燃过程相对缓慢；独立火焰面的重燃情节对应着较高的自由基爆炸指数和较小的温度梯度,且重燃过程相对较快。分析是为了更加准确、高效的模拟湍流燃烧现象。在不同的湍流扩散燃烧区内,主导因素差异巨大,这也对反应源项的封闭模型产生不同的要求。因此,本文提出基于实时模拟信息反馈的湍流扩散火焰模拟方法。该方法利用实时模拟数据进行分析,并反馈诊断信息到主程序的计算中：在积分反应源项时,根据分析的结果选择合适的模型进行封闭。实时反馈的信息为局部Damkohler数,可由化学爆炸模式分析(CEMA)和局部标量耗散率得到。依据反馈的Damkohler数,可以确定具体的湍流扩散火焰燃烧区,进而根据各个燃烧区的特点选择恰当的反应源项封闭模型：位于火焰面燃烧区时(Da≥DaLFA),可以使用基于化学制表的稳态层流火焰面模型进行封闭；位于熄火区时(Da≤1),此时采用基于Arrhenius方程的有限速率模型进行封闭：位于非稳态燃烧区时(1DaDaLFA),化学时间尺度相对于混合时间尺度来说是不能忽略的,火焰面模型也不再适用。本文在ODT程序的基础上嵌入了基于实时模拟信息反馈的模拟程序,并通过与原始ODT程序结果及实验值的对比,验证了基于实时信息反馈混合封闭模型的正确性。随后,本文又从湍流燃烧低维流形理论的角度对混合封闭模型进行了优化。本质上说,层流火焰面模型是一类基于假设条件的低维流形。本文摒弃了这种不利的前提条件,直接从高保真的模拟数据出发,构建了经验流形来替代稳态层流火焰面模型。由于直接来源于燃烧数据的分析,使得经验流形也能够对非稳态湍流燃烧区进行封闭；也使得基于经验流形的混合封闭模型对自由基组分的预测精度大大提高。此外,在构建经验流形时可以使用单次ODT模拟数据,且构建的经验流形对射流雷诺数、临界Damkohler数等条件依赖性较小。这些优点进一步扩大了基于实时信息反馈混合封闭模型的使用范围。本文主要的创新点如下：构建了基于Damkohler数和拉格朗日示踪方法的高精度计算燃烧分析平台,并据此鉴定、区分了不同情节的湍流扩散火焰的局部熄火重燃现象及其机理；构建了基于实时模拟信息反馈的湍流扩散火焰混合封闭模型,并从经验低维流形的角度进一步优化了混合封闭模型,通过与实验值及原始ODT程序的对比,验证了混合封闭模型的有效性。
[Abstract]:In essence, fire is a turbulent combustion phenomenon. The core problem of turbulent combustion is to study the interaction between turbulent mixing and complex chemistry at multiple time and space scales. The study of the interaction mechanism of turbulent mixing and complex chemistry helps to understand the deeper level of fire dynamics and help fire in special conditions. The development of the disaster model is to study the interaction of turbulent mixing and complex chemistry. First, it needs mass, including the reaction flow data of the dissipative scale turbulence information. In this paper, the one-dimensional turbulence model (ODT) is used to obtain the high fidelity data set.ODT model to distinguish the turbulence fluctuation information on the Kolmogorov scale, and in the coupled complex chemical calculation. Because the chemical reaction flow is a typical variable density flow, based on the original ODT program, the variable density version ODT program is developed on the basis of the original ODT program, considering the influence of the multidimensional velocity component and the density change on the random vortex event, and the effectiveness of the variable density ODT program is verified by comparison with the experimental data. The high precision ODT numerical platform which can distinguish the dissipative scale turbulence information provides a robust database for further analysis of the turbulent chemical interaction relationship and the construction of the empirical low dimensional manifolds. The study of the interaction of turbulent mixing and complex chemistry also requires efficient and accurate analysis of the reaction flow data. This paper is built on the basis of this paper. The Damkohler number and the Lagrange tracer method are extended to calculate the high precision combustion analysis platform. From the time scale relation, the interaction relationship between turbulent mixing and chemical reaction is decoupled. In this paper, the turbulent diffusion combustion process of CO syngas is numerically simulated by ODT program, and the theory of chemical explosion mode analysis (CEMA) is established and the theory of chemical explosion mode analysis is established. Based on the Damkohler number, the partial flameout plot analysis method of the turbulent diffusion flame is based on the Damkohler number. The results show that the quenching zone can be expressed by the large and negative Damkohler numbers. In addition, the explosion index of the Lagrange tracer unit is discussed in depth from the time evolution point of view. By distinguishing the duration of the flameout, two kinds of different kinds are defined. Two different turbulent diffusion flame reburning mechanisms are defined by the analysis of the explosion exponent of the reburning process. The premixed flame propagation plot corresponds to a long time preheating process and a higher temperature explosion index, and the reignition process is relatively slow; the heavy burning of the independent flame surface is the same. The section corresponds to the higher free radical explosion index and the smaller temperature gradient, and the reignition process is relatively fast. The analysis is to be more accurate and efficient in simulating the turbulent combustion phenomenon. In the different turbulent diffusion combustion regions, the dominant factors vary greatly. This also produces different requirements for the sealing model of the source terms. Therefore, this paper puts forward the basis of this paper. The method of turbulent diffusion flame simulation in real time simulated information feedback. This method uses real time analog data to analyze and feedback the diagnosis information to the calculation of the main program: when integrating the source term of the response source, the appropriate model is selected according to the results of the analysis. The information of real time feedback is a local Damkohler number, which can be made by a chemical explosion model. The CEMA and the local scalar dissipation rate are obtained. According to the Damkohler number of the feedback, the specific turbulent diffusion flame combustion zone can be determined, and then the appropriate source term closure model is chosen according to the characteristics of each combustion region: at the flame area (Da > DaLFA), the steady-state laminar flame surface model based on chemical tabulation can be used. The type is closed; at the time of the flameout area (Da < 1), the finite rate model based on the Arrhenius equation is used at this time to be closed: in the unsteady combustion zone (1DaDaLFA), the chemical time scale can not be ignored with respect to the mixing time scale, and the flame surface model is no longer applicable. This paper is embedded on the basis of the ODT program. The simulation program of information feedback is simulated in real time, and the correctness of the hybrid closed model based on real time information feedback is verified by comparison with the original ODT program results and experimental values. Then, this paper also optimizes the mixed closed model from the angle of turbulent combustion low dimensional manifold theory. In essence, laminar flame surface model is a class of models. In this paper, based on hypothetical low dimensional manifolds, this paper discarded this unfavorable condition and constructed an empirical manifold to replace the steady state laminar flame surface model directly from the high fidelity simulation data. The empirical manifolds can also be closed to the unsteady turbulent combustion zone because of the direct source of the combustion data. The prediction accuracy of the free radical component is greatly improved by the mixed closure model of the empirical manifold. In addition, the single ODT simulation data can be used in the construction of the empirical manifold, and the established empirical manifolds are less dependent on the Reynolds number and the critical Damkohler number. These advantages extend the mixed closure based on the real-time information feedback. The main innovation points of this paper are as follows: the high precision calculation combustion analysis platform based on Damkohler number and Lagrange tracer method is constructed, and the local quenching reburning phenomenon of turbulent diffusion flames in different plots and its mechanism are identified by this method, and the turbulent expansion based on real time simulation information feedback is constructed. The mixed closed model is optimized and the mixed closure model is further optimized from the angle of empirical low dimensional manifolds. The effectiveness of the hybrid closed model is verified by comparison with the experimental values and the original ODT program.
【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：O357.5

【相似文献】