End-of-Injection Effects on Diesel Spray Combustion
Knox, Benjamin W
MetadataShow full item record
Increasingly stringent emissions regulations have created a demand for cleaner burning engines. Low-temperature combustion (LTC) strategies have been proposed to meet low soot and nitrogen oxides emissions but LTC strategies suffer from excessive unburned hydrocarbon (UHC) and carbon monoxide (CO) emissions. These emissions have been shown to originate from overly fuel-lean mixtures near the nozzle that do not burn to completion. These mixtures are said to be over-mixed beyond a flammability limit and are caused by increased entrainment during end-of-injection. The coupling between end-of-injection entrainment and incomplete combustion of near-nozzle mixtures is not well understood, however, in part due to the large parameter space in engines. Thus, this thesis aims to develop tools and models to measure end-of-injection combustion observables and predict the likelihood of UHC and CO emissions over a wide range of conditions. This thesis seeks to perturb the coupling between end-of-injection entrainment and incomplete combustion of near-nozzle mixtures by systematically varying the ambient thermodynamic conditions, injection parameters, as well as the end-of-injection transient. Four distinct behaviors of the spray flame following end-of-injection were identified: soot recession, complete combustion recession, partial combustion recession, and no/weak combustion recession. Combustion recession is the process whereby the initially lifted reaction zone retreats back towards the nozzle immediately following end-of-injection, thus consuming UHC/CO that would otherwise remain near the nozzle. Soot recession spatially and temporally overlaps with combustion recession and is the result of igniting rich mixtures. Regression of a comprehensive dataset indicates that combustion recession is promoted with higher ambient temperatures, higher ambient oxygen concentrations, higher ambient densities, longer end-of-injection transients, lower injection pressures, and larger nozzle orifice diameters. Similar trends are observed for soot recession as well. Rather than rely solely on regression for predictions of combustion recession, a first-principles based approach was used to develop a scaling law for combustion recession that is applicable to a wider class of injectors and injection strategies than those tested experimentally. Using a definition for the local Damkohler number throughout the jet, a limiting location of ignition was identified and linked to the flame lift-off length to develop both an end-of-injection ignition timescale and a steady injection ignition timescale. The proportionality between the two timescales was used to predict the likelihood of combustion recession and thus UHC/CO emissions.