Autoignition of an n-heptane jet in a confined turbulent hot coflow of air

Abstract

The autoignition of a continuous, single jet of pure liquid n-heptane injected concentrically and axisymmetrically from a water-cooled circular nozzle into a confined turbulent hot coflow (CTHC) of air at atmospheric pressure has been investigated experimentally at air temperatures up to 1150 K and velocities up to 40 m/s. The aim of this work was to examine the emergence of liquid-fuel autoignition in the presence of flow, mixture and phase inhomogeneities, to which end, the velocity, temperature and fuel-droplet fields inside the CTHC reactor were characterized in a series of dedicated measurement campaigns. Distinct phenomena were identified concerning the emergence of various regimes: no autoignition, random spots, and continuous flame. In the random spots regime, autoignition appeared in the form of well-defined, discrete localized spots occurring randomly within the reactor, similar to observations in a similar apparatus with gaseous fuels (Markides, 2005; Markides and Mastorakos, 2005, 2011; Markides et al., 2007). High-speed optical measurements of these random spots were made from which the autoignition locations/lengths were measured, and then used to infer average autoignition delay, or residence, times from injection based on the bulk air velocity. An increase in the air temperature moved the region of autoigniting spots closer to the injector nozzle, thus decreasing the autoignition length and also decreasing the autoignition delay time. Generally, autoignition moved downstream with increasing bulk air velocity, but the delay times decreased contrary to the aforementioned earlier work with pre-vaporized n-heptane in this geometry. Of interest is the finding that at the highest investigated air velocities, the autoignition length decreased as the air velocity increased, which again deviates from the same earlier work with vaporized n-heptane. Furthermore, higher liquid injection velocities also resulted in increased autoignition lengths and times. The results from this work suggest that the turbulent inhomogeneous autoignition of liquid fuels cannot be directly predicted from chemical delay times of equivalent homogeneous systems, or even from knowledge of the turbulent inhomogeneous autoignition of single-phase (gaseous) fuels in the same reactor system (flow geometry, fuel, conditions, etc.), and that the effects of evaporation and turbulent mixing must be accounted for. The data can assist the development of advanced turbulent and multiphase reacting flow models.