Numerical simulation of flame propagation and localized preflame autoignition in enclosures
Highlights
► Novel approach for numerical simulation of confined explosions is developed. ► It is applied to model turbulent premixed combustion in vessels of different geometry. ► It allows continuous monitoring of both flame front and preflame reactions.
Introduction
Numerical simulation of flame propagation with preflame autoignition in enclosures is a complex problem. The phenomenology of the process includes flame ignition and propagation, unburned mixture compression and heating, as well as formation of hot spots and fast-spreading localized explosions in the preflame region. The localized explosions evolve from the sites with the minimum induction time and traverse the preflame zone as spontaneous ignition waves with the propagation velocity depending on the local instantaneous distributions of temperature and mixture composition. Better understanding of these phenomena is important for the improvement of existing measures aimed at preventing violent accidental explosions in process plants.
The objective of any combustion model in a CFD code is to provide correct values of mean reaction rates in each computational cell regardless the combustion mode (premixed, nonpremixed, partially premixed, homogeneous, inhomogeneous, spontaneous, frontal, etc.). The correct value of the mean reaction rate in the computational cell can be obtained only if one knows the reaction kinetics and instantaneous fields of temperature and species concentrations inside the cell. The development of reaction kinetics is the separate task which is independent of the CFD combustion modeling. The only relevant issue is the CPU time required for calculating instantaneous reaction rates. This issue can be overcome by applying properly validated short overall reaction mechanisms or look-up tables.
The instantaneous fields of temperature and species concentrations inside the cell are usually not known. Therefore, one has to replace this lacking information by combustion models. There exist many combustion models both for laminar and turbulent flows. If combustion chemistry is fast as compared to mixing, the Spalding (1976) Eddy-Break-Up model can be used. It is simple but has a limited range of validity. There is a whole class of statistical combustion models (based on the formalism of probability density functions (PDF)) with probabilistic representation of turbulence and its interaction with chemistry, Pope (1990). This approach is very attractive for treating both flame propagation and autoignition problems, however requires large CPU resources. The other class of models deals with a flamelet approach, Peters (1986). In this approach, the instantaneous flame is assumed to consist of localized reactive sheets, which are transported by the flow and wrinkled by turbulent eddies. The flamelet approach is applicable when the characteristic turbulent scales are larger than a typical flame thickness. This condition is satisfied in many practical situations.
The approach supposed below is a sort of combination of flamelet and statistical combustion models to allow for simultaneous treatment of frontal combustion by explicit tracing of mean reactive surfaces and volumetric combustion by the transported PDF approach. The availability of such an approach makes it possible to attack the problem of flame propagation with preflame autoignition in enclosures.
Section snippets
Flame-tracking method
The FT method deals with the subgrid model of laminar/turbulent combustion. The essence of the model can be readily explained on the example of laminar flame propagation. In the FT method, the flame-surface shape and area are found based on the Huygens principle: Each elementary portion of the flame surface displaces in time due to burning of the fresh mixture at local velocity un (normal to the flame surface) and due to convective motion of the mixture at local velocity V. The local
Particle method
The preflame zone exhibits volumetric reactions of fuel oxidation, formation of intermediate products like alcohols, aldehydes, peroxides, etc. In general, preflame reactions are inhomogeneous due to inhomogeneous distributions of temperature and main species concentrations and due to high sensitivity of reaction rates to these parameters. Therefore preflame reactions can result in localized energy release.
Direct (and CPU time consuming) way to calculate volumetric reaction rates is to solve
Results and discussion
Described below are the results of application of the FTP method to several test cases illustrating method capabilities.
Concluding remarks
A novel computational approach based on the coupled 3D FTP method has been developed and used for numerical simulation of confined explosions with/without preflame autoignition. The method allows continuous monitoring of both flame front and preflame reactions with generation of high-intensity secondary pressure peaks and is applicable to simulations of accidental explosions in various industrial reactors with complex geometry.
Acknowledgments
This work was partly supported by the Russian Foundation for Basic Research (grant 11-08-01297) and AVL LIST GmbH.
References (19)
Turbulent premixed combustion: flamelet structure and its effect on turbulent burning velocities
Progress in Energy and Combustion Science
(2008)- et al.
Overall kinetic mechanisms for modeling multistage autoignition of hydrocarbons in reacting flows
Russian Journal of Chemical Physics
(2006) - et al.
Kinetics of ‘blue’ flames in the gas-phase oxidation and combustion of hydrocarbons and their derivatives
Russian Chemical Reviews
(2007) - et al.
Data base for characteristics of n-heptane laminar combustion
How fast can we burn?
Proceedings of the Combustion Institute
(1992)Der einfuss der turbulenz auf die flammengeschwindigkeit in gasgemischen
Zs Electrochemie
(1940)- et al.
Combined flame tracking–particle method for numerical simulation of deflagration-to-detonation transition
Turbulent premixed flame propagation models for different combustion engines
Proceedings of the Combustion Institute
(1990)- et al.
Mathematical modeling of flame propagation in smooth tubes and in tubes with regular obstacles
Journal of Fire and Explosion Safety
(2010)
Cited by (15)
Updated conceptual design of hydrogen/ethylene fueled detonation ramjet: Test fires at Mach 1.5, 2.0, and 2.5
2022, Aerospace Science and TechnologyCitation Excerpt :The calculation technique is described in [26] and is not reproduced here. Briefly, in the calculations based on the Reynolds-Averaged Navier–Stokes equations, both the finite rates of molecular mixing of fuel components (hydrogen and air) and finite rates of chemical reactions in a multicomponent gas are considered by applying the particle method [56,57]. The calculation technique is earlier thoroughly validated against experimental data [37,58,59].
Air-breathing pulsed detonation thrust module: Numerical simulations and firing tests
2019, Aerospace Science and TechnologyCitation Excerpt :The effect of the dimensions of the computational cells on the flow structure was checked by additional calculations on essentially smaller meshes. Some examples of numerical solutions for the homogeneous mixtures can be found in [30,40]. Calculations were performed for the conditions when the TM is placed in a free air stream with the approaching flow velocity of 10 m/s with the injection of liquid kerosene into the combustor.
3D simulation of hydrogen ignition in a rapid compression machine
2013, Journal of Loss Prevention in the Process IndustriesCitation Excerpt :The chemical source terms are obtained using implicit numerical algorithm with internal time stepping to keep the CFL number here below 1. The combined algorithm has been thoroughly validated against experimental data on flame acceleration and deflagration-to-detonation transition in tubes with and without obstacles, Ivanov and Frolov (2011), as well as for shock-induced ignition and preflame autoignition in enclosures, Frolov, Ivanov, et al., 2013. The CFD software platform used for the numerical solution is AVL FIRE 2011.
Crankcase Explosions in Marine Diesel Engines: A Computational Study of Unvented and Vented Explosions of Lubricating Oil Mist
2024, Journal of Marine Science and Engineering