We analyse an autoignition process in a two-phase flow in a temporally evolving mixing layer formed between streams of a cold liquid fuel (heptane at 300 K) and a hot oxidizer (air at 1000 K) flowing in opposite directions. We focus on the influence of a discretization method on the prediction of the autoignition time and evolution of the flame in its early development phase. We use a high-order code based on the 6th order compact difference method for the Navier–Stokes and continuity equations combined with the 2nd order Total Variation Diminishing (TVD) and 5th order Weighted Essentially Non-Oscillatory (WENO) schemes applied for the discretization of the advection terms in the scalar transport equations. The obtained results show that the autoignition time is more dependent on the discretization method than on the flow initial conditions, i.e., the Reynolds number and the initial turbulence intensity. In terms of mean values, the autoignition occurs approximately 15% earlier when the TVD scheme is used. In this case, the ignition phase characterizes a sharp peak in the temporal evolution of the maximum temperature. The observed differences are attributed to a more dissipative character of the TVD scheme. Its usage leads to a higher mean level of the fuel in the gaseous form and a smoother distribution of species resulting in a lower level of the scalar dissipation rate, which facilitates the autoignition process.
2
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW
The paper presents large eddy simulation (LES) study aiming at investigations of an influence of flow conditions on a spark ignition process in a two-phase shear dominated flow. Implicit LES approach is applied for the combustion modelling and the spark is modelled using the energy deposition model of Lacaze et al. [20]. We examine an impact of turbulence intensities and randomness of initial distributions of velocity fluctuations on a flame development during the spark duration and shortly after it is switched off. It is found that for a strong spark, as used in IC engines, the turbulence intensity has little effect on the ignition and flame kernel growth and no significant differences are seen even if the turbulence intensities differ four times. It is observed that weak turbulent structures cannot affect fast flame propagation mechanism and its development is conditioned by evaporation and rapid thermal expansion. In such regimes, the turbulence seems to be too weak to significantly alter the flame dynamics. It is found that at the initial stage of the flame development it grows toward the fuel-rich region and spread over the fuel-lean side only after the evaporated fuel diffuses and mixes with the oxidizer stream. The flame size and its shape turn out to be equally dependent on the initial distribution of the turbulence fluctuations and turbulence intensity.
3
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW
The paper presents the results of LES simulation of two different turbulent channels with inlet conditions corresponding to the Reynolds number Re =395. In both cases a varying pressure gradient was obtained by an adequate curvature of one of the walls. The first case is treated as a benchmark and is used to validate the numerical procedure. This case is characterized by the same cross-section area at the inlet and outlet and a bump of a smooth profile located on one of the walls designed to be identical to the one used in the experiment conducted at Laboratorie de Mecanique de Lille (LML) (Marquillie et al., 2008). The second case corresponds to the geometry which reproduces the real geometry of the turbomachinery test section of the Czestochowa University of Technology. The test section was created in such a way as to produce the pressure gradient which would correspond to the conditions present in the axial compressor blade channel. The shape of both channels produced initially favorable (FPG) and then adverse pressure gradients (APG), and in this way created conditions for boundary layer separation. Due to a reverse flow where the turbulence transport is dictated by the dynamics of the large-scale eddies such a case is well suited to demonstrate predictive features of the LES technique.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.