Busch, Stephen; Zha, Kan Z.; Miles, Paul C.; Warey, Alok; Pesce, Francesco; Peterson, Richard; Vassallo, Alberto
A pilot-main injection strategy is investigated for a part-load operating point in a single cylinder optical Diesel engine. As the energizing dwell between the pilot and main injections decreases below 200 μs, combustion noise reaches a minimum and a reduction of 3 dB is possible. This decrease in combustion noise is achieved without increased pollutant emissions. Injection schedules employed in the engine are analyzed with an injection analyzer to provide injection rates for each dwell tested. Two distinct injection events are observed even at the shortest dwell tested; rate shaping of the main injection occurs as the dwell is adjusted. High-speed elastic scattering imaging of liquid fuel is performed in the engine to examine initial liquid penetration rates. The penetration rate data provide evidence that rate shaping of the initial phase of the main injection is occurring in the engine and that this rate shaping is largely consistent with the injection rate data, but the results demonstrate that these changes are not responsible for the observed trend in combustion noise. A zero-dimensional model is created to investigate the causes of the observed combustion noise behavior. The trend in simulated combustion noise values agree well with the experimentally determined trend, which is associated with two main factors: relative changes in combustion phasing of the pilot and main heat release events and suppression of the pilot apparent heat release for dwell times near the minimum-noise dwell. Two possible mechanisms by which the relative phasing between the pilot and the main heat release events impacts combustion noise are proposed.
Based on the ensemble-averaged velocity results, flow asymmetry characterized by the swirl center offset and the associated tilting of the vortex axis is quantified. The observed vertical tilting of swirl center axis is similar for tested swirl ratios (2.2 and 3.5), indicating that the details of the intake flows are not of primary importance to the late-compression mean flow asymmetry. Instead, the geometry of the piston pip likely impacts the flow asymmetry. The PIV results also confirm the numerically simulated flow asymmetry in the early and late compression stroke: at BDC, the swirl center is located closer to the exhaust valves for swirl-planes farther away from the fire deck; near TDC, the swirl center is located closer to the intake valves for swirl-planes farther away from the fire deck. It is evident from experimentally determined velocity fields that the transition between these two asymmetries has a different path for various swirl ratios, suggesting the influence of intake port flows. Flow field asymmetry can lead to an asymmetric mixture preparation in Diesel engines. To understand the evolution of this asymmetry, it is necessary to characterize the in-cylinder flow over the full compression stroke. Moreover, since bowl-in-piston cylinder geometries can substantially impact the in-cylinder flow, characterization of these flows requires the use of geometrically correct pistons. In this work, the flow has been visualized via a transparent piston top with a realistic bowl geometry, which causes severe experimental difficulties due to the spatial and temporal variation of the optical distortion. An advanced optical distortion correction method is described to allow reliable particle image velocimetry (PIV) measurements through the full compression stroke.
In this work computational and experimental approaches are combined to characterize in-cylinder flow structures and local flow field properties during operation of the Sandia 1.9L light-duty optical Diesel engine. A full computational model of the single-cylinder research engine was used that considers the complete intake and exhaust runners and plenums, as well as the adjustable throttling devices used in the experiments to obtain different swirl ratios. The in-cylinder flow predictions were validated against an extensive set of planar PIV measurements at different vertical locations in the combustion chamber for different swirl ratio configurations. Principal Component Analysis was used to characterize precession, tilting and eccentricity, and regional averages of the in-cylinder turbulence properties in the squish region and the piston bowl. Complete sweeps of the port throttle configurations were run to study their effects on the flow structure, together with their correlation with the swirl ratio. Significant deviations between the flows in the piston bowl and squish regions were observed. Piston bowl design, more than the swirl ratio, was identified to foster flow homogeneity between these two regions. Also, analysis of the port-induced flow showed that port geometry, more than different intake port mass flow ratios, can improve turbulence levels in-cylinder.
The rate at which fuel is injected into the cylinder of a direct injection Diesel engine has significant implications for the ensuing mixture formation and combustion processes. Advances in fuel injector technology enable a variety of advanced injection strategies, particularly very closely coupled injection events. In this work, a Moehwald HDA injection quantity and rate measuring unit is used to investigate the injection rates obtained with a pre-production solenoid injector with a fast acting, pressure-balanced control valve using a blend of n-hexadecane and heptamethylnonane (DPRF58). The effects of digital signal filtering on the rate shape and injected mass are investigated for a single injection. Additionally, the effects of physical parameters such as fuel and measurement chamber temperature, axial clamping force on the injector, high pressure line length, and solenoid current pull up time on the rate shape are investigated. The primary purpose of these simple parameter variations is to establish whether or not they have an impact on the measured injection rate traces and/or total measured injected masses. At each dwell time, the rates of injection are compared between the three injectors tested. These results show that these pre-production injectors can operate with very short dwell times while the injection rate curves indicate distinct pilot and main injection events and an influence of dwell on the rate shape of the main injection. Testing with PRF, a blend of n-heptane and isooctane, shows that while rates of injection are comparable to those obtained with the DPRF for a single injection, they are dramatically different for multiple injections. This has significant implications for the optical diagnostic techniques that may be employed to study the effects of multiple injections on the mixture formation process.