Previous Page  |  Table of Contents  |  Next Page

Sean C. Garrick, Principal Investigator

Large Scale Simulations of Turbulent Reacting Flows

This research is performing numerical simulations of turbulent reacting flows via the combination of large eddy simulation (LES) and reduced chemical kinetics. There are two major issues that must be resolved to facilitate LES of turbulent reacting flows-construction of an efficient numerical scheme for solving the governing transport equations and-implementation of a chemical kinetic mechanism that can accurately capture the chemistry of complex hydrocarbon flames. In this work, both direct numerical simulation (DNS) and LES are being performed. With this approach, the performance of the reduced mechanisms, and the numerical solution procedures can be assessed.

99/222

Traditionally, most contributions on DNS of turbulent reacting flows are via simplified (and somewhat naive) chemistry models (typically of the type Fuel + Air Æ Products). This is understandable since, with severe computational requirements of DNS, it is impossible, at least within the foreseeable future, to consider exact detailed chemistry mechanisms. However, development of computational procedures capable of dealing with turbulent flows involving realistic kinetic mechanisms is in order. The rigorous means of treating the influence of chemistry in hydrocarbon fuel combustion requires the consideration of detailed mechanisms of elementary reactions. The oxidation mechanism of fuels such as propane requires the consideration of on the order of one hundred reactions involving the transport of approximately thirty species for an accurate description of the flame structure. With the available computer power, such a consideration is very computation-intensive even for simple stead laminar flames. The situation becomes substantially worse when turbulent flames are considered. A remedy of this problem is to systematically reduce the number of equations characterizing the starting mechanism to a reduced system of equations that is more manageable. The research field of reduced kinetic schemes has been actively pursued by chemists for a long time. However, it has only been only since the 1980s that these schemes have been systematically utilized in combustion studies. Of course the number of steps taken in this approach must be prescribed in such a way that the essential physics of the problem is retained. For some more conventional hydrocarbon flames, the literature on reduced kinetic mechanisms is somewhat rich. Methane is one of the first fuels for which reduced schemes have been developed. In particular, four- and five-step mechanisms capable of simulating the mechanism of NOx formation are available. Reduced mechanisms are also available for other fuels (e.g., acetylene, ethylene, and propane).

A survey of chemical engineering and combustion literatures reveals relatively little work in LES of chemically reacting turbulent flows. It appears that Schumann was one of the first to conduct LES of a reacting flow. However, the assumption made in this work simply to neglect the contribution of the small scale chemical fluctuations is debatable. The importance of such fluctuations is well recognized in RANS computation of reacting flows in both chemical engineering and combustion problems. Therefore, it is natural to believe that these fluctuations will also have a significant influence in LES.

Hydrodynamic Stability in a Plane Jet (Vorticity Contours).

One approach that has proven particularly useful in accounting for the chemical fluctuations is based on the probability density function (PDF) or the joint PDF of the chemical species. This approach offers the advantage that all the statistical information pertaining to the chemical field is embedded within the PDF and any statistical quantity is readily at hand. The systematic approach for determining the PDF is arrived at by solving the transport equation governing its evolution. This is facilitated by introducing the filtered density function (FDF) and by providing an effective numerical means to simulate this FDF. Because of the added dimensionality of the compositional variables, solution of the FDF transport equation by conventional numerical methods is possible in only the simplest of cases. However, a Monte Carlo method may be used for this purpose. The use of these schemes have proven very effective in RANS. Considering the capabilities of the scheme and encouraging results in preliminary usage of this scheme, an element of this project is fully utilizing this methodology for LES of turbulent reactive flows.

The most notable differences between the Monte Carlo solution of the FDF as detailed above and the conventional (most current) procedures employed for LES include the fact the FDF accounts for the effects of chemical reactions in an exact manner regardless of the speed of reaction and the character of the flame zone (flamelet regime, distributed reaction zones, etc.), whereas in conventional LES procedures it is not at all clear how to model the unresolved reaction rate. The advantage of the FDF methodology over existing methodologies is the simple fact that the rate of chemical conversion appears exactly. There are no associated modeling issues, irregardless of how complex the chemical reaction mechanism. Such a robust methodology has applications in high-speed air-breathing propulsion, gas turbine engines, internal combustion engines, and biomass reactors just to name a few.

These researchers are using currently available reduced chemistry schemes in performing large eddy simulations of turbulent reacting flows. Based on preliminary assessments, they are implementing the reduced scheme of Seshadri and Peters. This mechanism is chosen primarily because it has been implemented in preliminary DNS studies. In particular, The Seshadri-Peters mechanism has predicted the presence of hydrogen radicals, H, in the flow field. The presence of H radicals is crucial for the reaction to progress as its absence causes the termination of reaction almost immediately. In addition, the mechanism facilitates the capturing of the flame structure. Such a detailed description cannot be obtained b using global one-step mechanisms.

With the implementation of such kinetics in a LES context, numerical combustion is one step closer to being a reality as an engineering tool.

This information is available in alternative formats upon request by individuals with disabilities. Please send email to alt-format@msi.umn.edu or call 612-624-0528.

HOME | QUESTIONS | FEEDBACK
Employment | Events | Links | People | Programs | Publications | Support | Welcome