Theoretical and Numerical Analysis of Laminar Ethanol Spray Flames for the creation of a Spray Flamelet Library

Abstract

Combustion of fuels appears in many sets of equipment and applications that involve transformation of a primary not useful form of energy into a secondary usable one that can be employed for purposes of electricity generation, transportation and industrial thermochemical processes. Strict environmental regulations demand for the reduction of the pollutants emissions (e.g. NO$_{x}$ CO$_{2}$, CO, SO$_{2}$, etc.) that are responsible for the global warming of the planet and the formation of acid rain and smog. Since renewable energy sources cannot replace the use of fossil fuels in the foreseeable future alternative ways have to be investigated. Combustion of biofuels such as liquid ethanol, a well known and widely used alcohol, under MILD (Moderate or Intense Low-oxygen Dilution) conditions is a scenario on that direction. From the one hand, combustion of biofuels offers sustainability and lower greenhouse gas emissions and from the other, MILD conditions offer more uniform and lower peak temperature profiles reducing so the nitrogen oxides emissions and increasing the lifetime of the combustion chamber. However, modeling of ethanol spray turbulent combustion is a complicated problem since it involves many different phenomena that are coupled between each other including the injection of the spray, the atomization into droplets and the dispersion in the domain, the turbulence of the gaseous flow, the mass, momentum and energy exchange between the two phases and finally the combustion. The closure of the unresolved chemical source terms in the numerical analysis of such a problem is a challenge and the Flamelet Generated Manifolds (FGM) method appears as a perspective approach to deal with it. This method approximates the turbulent flame as an ensemble of laminar, thin, one dimensional, counterflow flames the so-called flamelets. In that way the chemical reactions can be decoupled from the turbulence field and be calculated separately. The results are stored in look-up tables that are retrieved during the turbulence simulation with the use of connecting variables. Not having found a way so far to relate a spray flamelet look-up table with the turbulence flow many researchers create a gaseous flamelet table instead ignoring the spray and evaporation effects in it. As it has been reported this omittance can possibly lead to over-prediction of the temperature and important species profiles in the modeling of the turbulent flame. In this study, both a gaseous and spray ethanol flamelet library are created so that a direct comparison between the two can be done. The interference of spray influences the structure of the flame making it deviate significantly from the gaseous flamelets. The results show that not only the strain rate (as for the case of the gaseous flamelets), but also the equivalence ratio implied in the fuel side and the initial droplets’ diameter influence the flame structure of the spray flamelets. A non-slip velocity between the droplets and the carrier gas is kept as a boundary condition for all the cases. Increased initial droplets' momentum due to increased strain rate and initial diameter makes the droplets survive and follow longer trajectories in the domain. If droplets manage to cross the stagnation plane and meet the opposing gaseous stream, one or more droplets' reversals are observed. The motion of droplets in the domain is associated with the evaporation process the location of which affects the viability of the flame. Enhanced evaporation at the stagnation plane favors the creation of a double reaction zone. If the ethanol vapor is brought in the reaction zone at a high temperature level a hotter flame compared to the gaseous case can be created. Moreover, if evaporation takes place in regions where combustion is not strong enough, local flame extinction might happen. Thus, not only the residence time of the reactants (as for the case of the gaseous flamelets), but also the location and the intensity of the evaporation affect the flame structure. Decomposition of the ethanol gas into intermediate species can occur for all the cases but is more apparent in the gaseous flamelets and for lower strain rates. Mass fraction profiles of CO$_{2}$ and H$_{2}$O indicate the outer flame structure and follow the temperature profile, while local presence of CO indicates that the reactions are weakened, incomplete or they have just started. Local evaporation of the droplets is also predicted by the evaporation source term of the continuity equation and at the location where a local peak is observed additional peaks of the ethanol and oxygen consumption rates are found which are in a stoichiometric ratio. For increased droplets' initial momentum, slip velocity effects are found that lead to increased drag force and the deceleration of the droplets. Increasing the equivalence ratio at the boundary two reaction zones are observed for all the cases, but it does not influence the motion of the droplets. Finally, as a result of the evaporation two or more peaks of scalar dissipation rate can be found compared to the single peak observed in the gaseous flamelets. It becomes evident that the motion and the evaporation of the droplets strongly affects the flame structure. More reaction zones, of varying intensity and location can be found in the spray flamelets. Temperature and species profiles show deviations between the two libraries and as a result the creation of a complete spray flamelet library and finding a way to relate it with the turbulence calculation is an important issue to be studied in a future study. Finally, it is also recommended that the influences of a slip boundary velocity condition should be studied for the spray calculations.