Numerical Simulation Of Turbulent Diffusion Flame Based On Real-time Simulative Information Feedback

In essentially, fire is a turbulent combustion phenomenon. The key issue of turbulent combustion is the investigation of the interactions between turbulent mixing and detailed chemistry, which spans several orders of magnitude in spatial and temporal scales. The investigation of this interactions may contribute to improving our understanding in fire dynamics and developing new model for fire in special conditions.The investigation of the interactions between turbulent mixing and detailed chemistry need reactive fluid data at the Kolmogorov scale. In this work, One-dimensional turbulence (ODT) model is used to obtain these high fidelity data. The ODT model can capture the information at dissipation scale, and has an obvious cost advantage over multi-dimensional Direct Numerical Simulations. As reactive fluid is a variable-density fluid, an extension code for variable-density ODT is developed by original version. And the results of variable-density ODT are validated through quantitative comparison with experiment values. The ODT model, which can capture turbulent information at dissipation scale, provide robust data for empirical low-dimensional manifold and investigation of the turbulence-chemistry interactions.The investigation of the interactions between turbulent mixing and detailed chemistry also need an efficient and accurate analysis tool for reactive fluid data. From relative timescale, an analysis platform based on extension definition of Damkohler number and Lagrangian tracer method for calculation combustion is built to decouple the interactions between chemical reactions and turbulent flow. Then, an investigation of the local flame extinction and re-ignition in CO syngas non-premixed flame are carried out using ODT model. Combining with chemical explosive mode analysis (CEMA), diagnostic tools based on Damkohler number have been defined to distinguish the different scenarios of local extinction and re-ignition in turbulent diffusion flame. The results show that the extinction regime can be characterized by a large negative Damkohler number. In addiction, investigations of explosion index (El) for Lagrangian particles are performed. And two scenarios of local extinction are shown though extinction duration. Meanwhile, two re-ignition mechanisms are distinguished by the explosion index of the process prior to re-ignition. It is found that re-ignition mechanism through premixed flame propagation corresponds to a rather long preheat period, high temperature explosion index and slow re-ignition process. While re- ignition mechanism via independent flamelet corresponds to a high radical explosion index, small temperature gradient and rapid re-ignition process.The purpose for the analysis is to simulate turbulent combustion phenomena. In different turbulent diffusion combustion regimes, there are an obvious distinguishes in dominated factor for reactive mixture and closed model for reactive source terms. Therefore, a simulation method based on real-time simulative information feedback for turbulent diffusion flame is proposed. In this method, the real-time simulation data are analyzed, and the diagnosis information are returned to main program. Based on the analysis results, suitable closed models will be selected to calculate reactive source terms. The real-time feedback information are localized Damkohler numbers, which can be calculated by chemical explosive mode analysis (CEMA) and localized scalar dissipation rate. Based on the feedback Damkohler numbers, turbulent diffusion combustion regimes are distinguished, and suitable closed models are selected. For flamelet combustion regime (Da≥DaLFA),the steady laminar flamelet model is appropriate; for extinction regime (Da≤1), finite-rate chemistry based on Arrhenius equations should be considered; for unsteady regime (1<Da<DaLFA),the chemical timescale is non negligible compared mixing timescale, and the flamelet-like model is not suitable. In this work, the real-time simulative information feedback strategy is integrated in the ODT code, and is validated by quantitative comparison with experiment values and original ODT results.Then, a modified strategy is proposed for hybrid closed strategy from low-dimensional manifold of turbulent combustion. In essentially, the flamelet-like model is a typical low-dimensional manifold with some assumed conditions. In this work, the preconditions of flamelet-like model are given up, and empirical manifold obtained by high fidelity data is used to replace the steady laminar flamelet model. As empirical manifold is constructed based on samples of the compositions observed in turbulent combustion, it can be used to close the reactive source term in unsteady turbulent diffusion regime. And the simulation precision of radical based on empirical manifold is also enhanced greatly. In addiction, only single realization data of ODT are enough to construct empirical manifold, and jet Reynolds number and critical Damkohler number have little influence on empirical manifold. These advantages also extend the application range of the real-time information strategy based on empirical manifold.The main innovation can be summarized as follows. Firstly, an analysis platform based on Damkohler number and Lagrangian tracer method for calculation combustion is built. Different extinction and re-ignition scenarios and mechanisms are also investigated. Secondly, a hybrid closed strategy based on real-time simulative information feedback for turbulent diffusion flame is proposed, and then it is modified by empirical low-dimensional manifold. This strategy is also validated though quantitative comparison with experiment values and original ODT results.